[Senate Hearing 118-474]
[From the U.S. Government Publishing Office]
S. Hrg. 118-474
READING THE ROOM:
PREPARING WORKERS FOR AI
=======================================================================
HEARING
BEFORE THE
SUBCOMMITTEE ON EMPLOYMENT AND WORKPLACE
SAFETY
OF THE
COMMITTEE ON HEALTH, EDUCATION,
LABOR, AND PENSIONS
UNITED STATES SENATE
ONE HUNDRED EIGHTEENTH CONGRESS
SECOND SESSION
ON
EXAMINING PREPARING WORKERS FOR AI
__________
SEPTEMBER 25, 2024
__________
Printed for the use of the Committee on Health, Education, Labor, and
Pensions
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]
Available via the World Wide Web: http://www.govinfo.gov
__________
U.S. GOVERNMENT PUBLISHING OFFICE
57-256 PDF WASHINGTON : 2025
-----------------------------------------------------------------------------------
COMMITTEE ON HEALTH, EDUCATION, LABOR, AND PENSIONS
BERNIE SANDERS (I), Vermont, Chairman
PATTY MURRAY, Washington BILL CASSIDY, M.D., Louisiana,
ROBERT P. CASEY, JR., Pennsylvania Ranking Member
TAMMY BALDWIN, Wisconsin RAND PAUL, Kentucky
CHRISTOPHER S. MURPHY, Connecticut SUSAN M. COLLINS, Maine
TIM KAINE, Virginia LISA MURKOWSKI, Alaska
MAGGIE HASSAN, New Hampshire MIKE BRAUN, Indiana
TINA SMITH, Minnesota ROGER MARSHALL, M.D., Kansas
BEN RAY LUJAN, New Mexico MITT ROMNEY, Utah
JOHN HICKENLOOPER, Colorado TOMMY TUBERVILLE, Alabama
ED MARKEY, Massachusetts MARKWAYNE MULLIN, Oklahoma
TED BUDD, North Carolina
Warren Gunnels, Majority Staff Director
Bill Dauster, Majority Deputy Staff Director
Amanda Lincoln, Minority Staff Director
Danielle Janowski, Minority Deputy Staff Director
------
SUBCOMMITTEE ON EMPLOYMENT AND WORKPLACE SAFETY
JOHN HICKENLOOPER, Colorado, Chairman
ROBERT P. CASEY, JR., Pennsylvania MIKE BRAUN, Indiana, Ranking
TAMMY BALDWIN, Wisconsin Member
TIM KAINE, Virginia ROGER MARSHALL, M.D., Kansas
BEN RAY LUJAN, New Mexico MITT ROMNEY, Utah
ED MARKEY, Massachusetts TOMMY TUBERVILLE, Alabama
BERNIE SANDERS (I), Vermont, (ex TED BUDD, North Carolina
officio) BILL CASSIDY, M.D., Louisiana, (ex
officio)
C O N T E N T S
----------
STATEMENTS
WEDNESDAY, SEPTEMBER 25, 2024
Page
Committee Members
Hickenlooper, Hon. John, Chairman, Subcommittee on Employment and
Workplace Safety, Opening statement............................ 1
Braun, Hon. Mike, Ranking Member, U.S. Senator from the State of
Indiana, Opening statement..................................... 3
Witnesses
Meyer, Ken, Senior Director of Human Resources, Ryan Health, New
York, NY....................................................... 4
Prepared statement........................................... 5
Kimbrough, Dr., Karin, Chief Economist, Linked In, Sunnyvale, CA. 9
Prepared statement........................................... 11
Kotran, Alex, Chief Executive Officer, aiEDU, San Francisco, CA.. 16
Prepared statement........................................... 18
Wilson, Denzel, Grassroots Program Manager, Seed AI, Washington,
DC............................................................. 22
Prepared statement........................................... 24
ADDITIONAL MATERIAL
Statements, articles, publications, letters, etc.
Hickenlooper, Hon. John:
Statement of the AFL-CIO Technology Institute................ 46
Braun, Hon. Mike:
Letter from the National Retail Federation................... 48
Markey, Hon. Ed:
Letter from the National Nurses United....................... 49
QUESTIONS FOR THE RECORD
Response by Dr. Karin Kimbrough to questions of:
Senator Lujan................................................ 50
Response by Denzel Wilson to questions of:
Senator Lujan................................................ 51
READING THE ROOM:
PREPARING WORKERS FOR AI
----------
Wednesday, September 25, 2024
U.S. Senate,
Subcommittee on Employment and Workplace Safety,
Committee on Health, Education, Labor, and Pensions,
Washington, DC.
The Subcommittee met, pursuant to notice, at 2:01 p.m., in
room 562, Dirksen Senate Office Building, Hon. John
Hickenlooper, Chairman of the Subcommittee, presiding.
Present: Senators Hickenlooper [presiding], Casey, Kaine,
Markey, Braun, and Budd.
OPENING STATEMENT OF SENATOR HICKENLOOPER
Senator Hickenlooper. The Subcommittee on Employment and
Workplace Safety will come to order. We are here today to talk
about preparing workers across America to effectively use AI in
the workplace.
Ranking Member Braun and I will each give an opening
statement, then we will introduce the witnesses. After
witnesses give their testimony, each Senator, those of us here,
watching online, or coming into the meeting will each have 5
minutes to ask their questions.
Last year, this Subcommittee heard from witnesses on the
potential benefits of AI to our economy but concluded that
those benefits will only become a reality if we have a well-
trained workforce.
Since then, AI has only continued to explode with
innovation and has remained at the forefront of conversations
for both employers and workers, as new applications of AI
continue to transform the workplace.
Unlike most previous technologies--like all previous
technologies, previous technologies like personal computers or
cell phones, they initially had substantial barriers that
limited consumer access.
AI is already--has achieved wide access. AI powered
applications are being used by students and workers and
business owners all across the country. By some estimates, more
than 60 percent of companies are exploring how to integrate
some form of generative AI technology, even as we speak.
Additionally, some workers already have their own
subscriptions to AI applications and are using them to enhance
their work, to accelerate their projects. Bottom line, AI
clearly does have the potential to change how we all work.
While some tasks may become more automated, the majority of
jobs will use processes that employ in some form of
collaboration between AI and human run systems. Despite the
clear interest in AI technologies from employers and workers
alike, we have more work to do to create widely available AI
literacy training opportunities to put everyone on an even
playing field.
The rapidly changing landscape in AI technologies is making
some employers and even some industries hesitant to invest in
comprehensive training opportunities. They are not sure if what
they are training will be useful and remain relevant in a
relatively short period of time.
But we know that having a well-trained and informed
workforce is a key--it is really essential to making sure that
AI is used responsibly, and that both workers and businesses
can reap the full benefits of the tools.
For example, human talent is needed to evaluate outputs
generated by AI for accuracy or to tailor AI generated concepts
into customized solutions that support their customers. AI
literacy should also include an emphasis on methods to help
workers identify AI generated content versus human generated
content.
AI literacy training is going to help empower employers to
choose the safest applications for their workforce and make
sure workers can give well informed feedback about their
experience with AI.
I think we have read the room and now is the time to make
sure every worker has access to the professional development
training that they need to succeed. That is why earlier this
year, we introduced the Lifelong Learning Act with Senators
Budd and Peters to make sure we can appropriately invest in
training opportunities for current and future workers.
Senator Braun and I have also been working together to make
sure that the Department of Labor, as well as other agencies,
understand the urgent need for these programs and need to
address the safety factors connected to that urgent need.
During today's hearing, we will hear from panelists from
various sectors and communities who are providing or benefiting
AI literacy training opportunities. As AI technologies and
training programs change over time, we will need everyone, our
union partners, employers, nonprofits, everyone to make sure
that we get this right and we set ourselves up for success.
Before I yield to Senator Braun for his opening remarks, I
would like to enter into the record a statement from the AFL-
CIO Technology Institute about the importance of these
partnerships, the importance of workers' voices as we scale
training opportunities.
[The following information can be found on page 46 in
Additional Material:]
Senator Hickenlooper. With that, I will turn it over to
Ranking Member Braun for his opening remarks and to introduce
our first witness.
OPENING STATEMENT OF SENATOR BRAUN
Senator Braun. Thank you, Mr. Chairman. Thank you to the
witnesses for being here today. Since our Subcommittee's last
hearing on this subject matter, AI technology has only
advanced--its prevalence has only grown in the American
workplace.
Today, we are going to take a look at AI technology, how it
is affecting employers and employees, and learn what they
should be prepared for down the road. Prior to becoming a
Senator, I spent 37 years running a logistics and distribution
company and we poured technology into it.
There is no doubt that when you use technology correctly,
it gives you unbelievable benefits, but we have never seen this
horizon before. Embracing technology should be beneficial to
everyone. It generally has been.
But here, when the people that have brought it to the
forefront have issued it with a stark warning, be careful, we
would actually like it to be regulated. You have never heard
that in the past about most things that come to the
marketplace.
There is broad interest about this from across the
political spectrum, and certainly it shouldn't be politicized
along the way. If there is a role for Congress to play in
regulating this technology, it should come from following
careful consideration and informed recommendations from
nonpartisan experts in all affected fields.
It is going to take a different type of framework to fully
implement and regulate such advanced technology. However, in
doing so, the goal shouldn't be to kill the technology by
smothering it, because if we do that, we never get to the point
of how it is going to be beneficial.
The past couple of years we have seen unprecedented
development and adoption of it across industries. We must
ensure that Government leaders are trained to keep up with the
advancements in AI and recognize the benefits and risks of this
powerful tool.
In this place, and most governments, have been the slowest
to embrace normal technology along the way. So it is critical
that if we are going to try to do something to create a
framework, that we learn something about it.
Senator Gary Peters and I introduced the AI Leadership
Training Act to create an AI training program for Federal
supervisors and management officials. The aim of this bill is
to help improve the Federal workforce's understanding of its
applications and to ensure that leaders who oversee the use of
these tools understand its full potential benefits and risks.
Legislation like this is a small step to improve our
understanding of this technology and how it will impact our
workplaces in the immediate future. Today, we are going to hear
from witnesses who will shed light on some of those
expectations and will also provide examples of how AI is
already affecting the American workplace.
We should aspire to bring the technology to the point where
it can create benefits for the workplace, but we should also
continue to be mindful and have conversations about where this
technology can take us along the way. That is the purpose of
the hearing, and I look forward to it.
Before I yield, Mr. Chairman, I would like to submit a
letter from the National Retail Federation which outlines the
views of the retail industry on this important matter.
Senator Hickenlooper. Great. And were you going to
introduce the first witness, Mr. Meyer?
Senator Braun. After you say it is Okay to put this into
the record.
Senator Hickenlooper. Without objection.
[The following information can be found on page 48 in
Additional Material:]
Senator Braun. There we go. Okay. Thank you. My privilege
to introduce Mr. Ken Meyer. Mr. Meyer is a Senior Director of
Human Resources at Ryan Health, which is based in New York
City.
He is an HR professional with nearly 40 years of experience
and has served as a President of the New York City Chapter of
the Society for Human Resource Management. We welcome your
expertise here today. Thanks for being here.
STATEMENT OF KEN MEYER, SENIOR DIRECTOR OF HUMAN RESOURCES,
RYAN HEALTH, NEW YORK CITY, NY
Mr. Meyer. Thank you very much, Senator Braun. Good
afternoon, Chairman Hickenlooper and Ranking Member Braun.
Thank you for bringing us together for this important
conversation about preparing workers for AI.
My name is Ken Meyer, and I am the immediate past President
of the New York City Chapter of SHRM, the Society for Human
Resource Management. As the trusted authority on all things
work, SHRM as the foremost expert, researcher, advocate, and
thought leader on issues and innovations impacting today's
evolving workplaces.
With nearly 340,000 members in 180 countries, SHRM touches
the lives of more than 362 million workers and their families
globally. I have over 35 years of human resources experience
within the health care sector.
I currently serve as the Senior Director of Human Resources
at Ryan Health, a Manhattan based community health center with
nearly 500 employees. My testimony today is on behalf of SHRM
and does not necessarily reflect the views of Ryan Health.
SHRM's research shows that workplaces around the country are
already using AI to enhance their organization's capabilities.
One in four organizations currently use AI to support HR
related activities, with nearly two-thirds of these adopters
implementing AI in HR within the past year. A SHRM survey from
earlier this year finds that among HR leaders whose
organizations are currently using generative AI, 75 percent
report that it has enhanced efficiency, 69 percent say it has
increased creativity, and 65 percent say it has improved work
quality.
As more businesses embrace AI, American workers will need
opportunities to learn, adapt, and use this technology to
remain competitive. SHRM's research shows that 30 percent of HR
leaders already feel greater pressure to innovate, and 28
percent believe there is an increased need to upskill and
reskill workers.
Because of this, AI curiosity and AI literacy are new,
must-have skills for workers in nearly everywhere role.
Workers, including their managers, need to understand both the
opportunities and risks presented by AI.
This training helps managers identify when AI can enhance
efficiency and when it is inappropriate or too risky to use so
they can manage these risks. SHRM supports thoughtful
legislation and regulation that promotes rather than stifles
workplace and workforce innovation.
However, state and local legislative and regulatory efforts
regarding AI are creating a confusing patchwork of obligations
for employers, putting some AI applications out of reach of the
small and medium sized businesses because of the cost and
uncertainty of regulatory compliance.
Last year, New York City began imposing new requirements on
employer uses of automated employment decision tools. The cost
of complying with the new law and the uncertainty regarding AI
regulation mean that employers, especially small and medium
sized businesses, may choose not to use AI, which unfortunately
may place them at a competitive disadvantage.
SHRM believes that overlapping laws and regulations
regarding AI may lead to unintended consequences that create
uncertainty and discourage workplace innovation. SHRM supports
a uniform Federal standard that provides a clear framework for
how employers should strive to prevent unlawful bias when using
AI.
We also support Federal efforts to educate all stakeholders
about the benefits and risks of AI in the workplace. As the
trusted thought leader on all things work, SHRM is committed to
helping employers responsibly navigate this new landscape,
ensuring that AI enhances the workplace without compromising
fairness, inclusion, and diversity.
We also stand ready to provide Congress with our expertise
on workplace issues to ensure that you are fully informed about
the potential consequences of forthcoming public policy and
legislative efforts.
Thank you again for this opportunity to speak today and I
look forward to hearing your questions.
[The prepared statement of Mr. Meyer follows.]
prepared statement of kenneth meyer
Introduction:
Chairman Hickenlooper, Ranking Member Braun, and distinguished
Members of the HELP Subcommittee on Employment and Workplace Safety,
thank you for bringing us together for this important conversation
about the impact of artificial intelligence (AI) on the workforce. My
name is Ken Meyer, and I had the honor in 2023 of serving as President
of the New York City chapter of the Society for Human Resource
Management (SHRM). Today, I am here to share insights on behalf of
SHRM, the foremost expert, researcher, advocate, and thought leader on
issues impacting today's evolving workplaces.
As the trusted authority on all things work, SHRM is the foremost
expert, researcher, advocate, and thought leader on issues and
innovations impacting today's evolving workplaces. With nearly 340,000
members in 180 countries, SHRM touches the lives of more than 362
million workers and their families globally.
I serve as Senior Director for HR for Ryan Health, a Manhattan-
based community nonprofit healthcare provider with nearly 500
employees. I have over 35 years of human resource (HR) experience
within the healthcare industry. My written testimony is on behalf of
SHRM and does not necessarily reflect the views of Ryan Health.
The Intersection of AI and HR: The Benefits:
SHRM research \1\ shows that 1 in 4 organizations (26 percent)
currently use AI to support HR-related activities, with nearly two-
thirds of these adopters only implementing AI in HR within the past
year. As the architects of talent and performance management, HR
professionals are embracing AI to enhance their organizations'
capabilities because AI offers a promising array of solutions to
address the complex demands of the modern workplace. The top three
areas where organizations are already using AI to support HR-related
activities include recruitment, interviewing, and hiring (64 percent),
learning and development (43 percent), and performance management (25
percent).
---------------------------------------------------------------------------
\1\ SHRM's 2024 Talent Trends Survey was conducted in January 2024
and collected responses from over 2,300 HR professionals representing
organizations of all sizes and across a variety of industries in the
United States. A summary excerpt of these research findings is
available at https://shrm-res.cloudinary.com/image/upload/AI/2024-
Talent-Trends-Survey--Artificial-Intelligence-Findings.pdf.
Furthermore, the steady proliferation and integration of AI
underscores the need for organizations to enhance AI curiosity and
literacy to get the best return on investment. Another SHRM survey \2\
conducted in January and February 2024 finds that among HR leaders
whose organizations are currently using generative AI, 75 percent
report enhanced organizational efficiency, 69 percent say it has
increased creativity, and 65 percent say it has improved work quality.
SHRM's AI in the Workplace found U.S. workers recognize the need for
various skills or competencies to effectively collaborate with AI
technologies. U.S. workers identified ``technical skills to navigate
interfaces and tools'' (72 percent), ``digital literacy'' (68 percent),
and ``critical thinking skills'' (68 percent) as the top three
competencies needed to effectively collaborate with AI technologies in
the workplace.
---------------------------------------------------------------------------
\2\ SHRM conducted additional research, collecting 1,220 responses
from HR leaders completed between January 25, 2024, and February 7,
2024. The survey was fielded electronically using the SHRM Voice of
Work Panel to U.S.-based HR leaders who hold a title of Director or
higher and engage in HR activities. A copy of an article further
summarizing this research is available at https://www.shrm.org/
executive-network/insights/new-shrm-research-shows-how-hr-leaders-
really-feel-about-ai.
Three in four HR professionals believe that advancements in AI will
increase the importance of human intelligence in the workplace over the
next 5 years. SHRM believes it is essential to approach AI's adoption
responsibly, and we are helping our members address ethical
considerations, ensure transparency, and provide appropriate training.
All are crucial aspects of a successful AI integration strategy that
helps to reduce the risk of algorithmic discrimination.
AI + HI [Human Intelligence] Equals ROI [Return on Investment]
As organizations eagerly adopt AI technologies, grasping their
potential to create value will be essential for making informed
decisions. To successfully navigate workforce changes during this
significant technological transformation, individuals at all levels of
the workforce must learn how to \3\:
---------------------------------------------------------------------------
\3\ AI in the Workplace Playbook, SHRM, 2024
` Manage disruption and empower innovation. Work is where
markets, people, and disruptive technologies intersect with the
challenges of human intelligence. As generative AI becomes more
integrated into our daily lives, we will learn how to unlock
potential, spark innovation, and discern our unique human
---------------------------------------------------------------------------
qualities from AI.
` Augment the power of people to drive the world forward. To
flourish in the AI era, we must rethink work, workers, and
workplaces--reimagining a world of work where human
intelligence and ingenuity are upskilled, not replaced.
` Engender a culture of change and invest in human catalysts.
Studies show human catalysts are key to successful AI
integrations. Organizations must invest in their people to
empower the mindsets, skill sets, and toolsets to drive
thriving, responsible workplace transitions.
As SHRM President and Chief Executive Officer Johnny C. Taylor, Jr.
has said, ``The opportunities AI presents are limitless. When combined
with human ingenuity (HI), this synergy has the capacity to transform
your organization while maximizing human potential. AI + HI = ROI.''
SHRM believes that while AI can help augment and automate routine
tasks, we will reach our fullest potential by combining AI with human
intelligence and oversight. Through this combination of technology with
human intelligence, HR professionals are achieving positive outcomes.
Below are examples of how SHRM's members are combining AI with HI
across a wide range of HR applications and industries:
``We were having a hard time filling lower-skill-level
positions [ . . . ] due to the requirement that these workers
needed a high school diploma. By having AI scan for experience
instead of just a high school diploma, we were able to increase
our applicant pool and fill positions we were having a hard
time retaining. Our retention level in these jobs has risen,
and we have eliminated the requirement of a high school diploma
for these positions.''
--Large employer in the Health Care industry
``Recruiting is one instance. AI may identify a passive
candidate that we were unaware of. We have chosen to interview
and hire someone that has come to our notice this way.''
--Large employer in the Retail industry
``Chatbots are creating 24/7 access to answers, which improves
the candidate experience and frees up the recruiter to actually
recruit.''
--Large employer in the Child Care, Community, or Social
Services industry
``AI has provided us with ways to deepen our outreach when
recruiting, which, combined with the expertise of our
recruiters, has allowed us to find candidates we might have
never reached previously.''
--Small employer in the Professional and Civic Associations
industry
``We use AI to recommend learning pathways for our staff based
upon the types of trainings that they are watching/learning
from. We use human intelligence to conduct a gap analysis of
what might be crucial learning that is overlooked because of
how the AI provides pathways for people and then find ways to
push those overlooked training topics to relevant people.''
--Extra-large employer in the Professional, Technical, and
Scientific Services industry
``We've utilized AI for comprehensive job descriptions and to
create a defined role that provides clarity and structure
within the organization. This has helped new hires understand
the position and expectations and has also allowed us to
promote individuals with clear intent on their expanding
function.''
--Large employer in the Real Estate industry
``We integrated our [human capital management] software with an
AI-based engagement tool. It uses AI to recommend ways to
improve our engagement score in areas that are low. This is
being used by managers once they get their results to implement
change, based off of AI recommendations, to improve our
scores.''
--Extra-large employer in the Manufacturing industry
The Increased Demand for AI Literacy and Upskilling
As more businesses embrace AI, American workers will need
opportunities to learn, adapt, and use AI to remain competitive in the
global job market. SHRM's research \4\ shows that 30 percent of HR
leaders already feel greater pressure to innovate, and 28 percent
believe there is an increased need to upskill and reskill workers. AI
curiosity and AI literacy are becoming essential skills for employees
in nearly every role in the workforce.
---------------------------------------------------------------------------
\4\ SHRM conducted additional research, collecting 1,220 responses
from HR leaders completed between January 25, 2024, and February 7,
2024. The survey was fielded electronically using the SHRM Voice of
Work Panel to U.S. based HR leaders who hold a title of Director or
higher and engage in HR activities. A copy of an article further
summarizing this research is available at https://www.shrm.org/
executive-network/insights/new-shrm-research-shows-how-hr-leaders-
really-feel-about-ai.
` The increased demand for AI literacy and upskilling is
critical to the modern workforce. As AI continues to transform
workplaces, it is essential for employees to develop the skills
needed to work alongside these technologies. A SHRM Current
Events Pulse \5\ survey in August 2024 revealed that most
workers are still at the start of their AI journeys. For
example:
---------------------------------------------------------------------------
\5\ August 2024 Current Events Pulse, SHRM, 2024
Y 80 percent of U.S. workers classify their
understanding of AI as either beginner or intermediate,
while only 20 percent say they are at an advanced or
---------------------------------------------------------------------------
expert level.
Y Nearly a quarter of U.S. workers (22 percent) lack
any experience with AI, while 63 percent classify their
AI proficiency as beginner or intermediate. Only 15
percent of workers say they have advanced or expert-
level experience with AI.
Moreover, SHRM's research shows that the increased demand for AI
literacy and upskilling is disproportionately borne by older
generations, who are significantly more likely to say they are at a
beginner or intermediate level of AI understanding compared to younger
workers.
By investing in training programs and promoting AI literacy,
organizations can ensure that their workforce remains competitive and
can leverage AI to drive workplace innovation and productivity.
To address the growing need for AI literacy, many companies are
investing in upskilling and reskilling initiatives, using the
technology to create personalized learning and development
opportunities. These programs are designed to help employees develop
the skills they need to work alongside AI technologies. Training helps
managers identify when generative AI can enhance efficiency and when it
is inappropriate or too risky to use.
SHRM has been actively preparing employers and HR professionals to
navigate the implementation of AI in the workplace. By providing
comprehensive resources and expert guidance, SHRM helps organizations
understand the potential of AI to enhance productivity and innovation.
This includes:
` Developing a robust toolkit for using AI in employment: \6\
SHRM offers resources that help HR professionals understand how
to integrate AI into various HR functions, such as talent
acquisition, education and development, employee engagement,
and performance management.
---------------------------------------------------------------------------
\6\ https://www.shrm.org/topics-tools/tools/toolkits/using-
artificial-intelligence-employment-purposes.
` Creating a comprehensive AI specialty credential: \7\ This
6-week training program helps HR professionals learn to utilize
AI and create an AI implementation plan.
---------------------------------------------------------------------------
\7\ https://www.shrm.org/topics-tools/news/organizational-
employee-development/shrm-ai--hi-specialty-credential--ushering-in-a-
new-era.
` Collaborating with industry experts: SHRM works with experts
to develop best practices for integrating AI into HR functions,
ensuring that organizations can harness the benefits of AI
while mitigating risks.
Legislative and Regulatory Considerations:
SHRM is working to educate its members regarding both the benefits
and the risks of AI, and steps that should be taken to mitigate these
risks. SHRM supports thoughtful legislation and regulation that
promotes, rather than stifles, workplace and workforce innovation.
State and local legislative and regulatory efforts regarding AI are
creating a patchwork of obligations for employers, putting some AI
applications out of reach for small and medium-sized businesses due to
the cost and uncertainty of regulatory compliance.
For example, small and medium-sized organizations operating in New
York City are navigating unique regulatory environments. Last year, New
York City Local Law 144 became effective, imposing new requirements
when an employer uses ``automated employment decision tools.'' The cost
of complying with New York City's new law and the uncertainty regarding
AI regulation mean that small and medium-sized businesses may avoid
leveraging AI's potential until there is stability, placing them at a
competitive disadvantage.
SHRM is concerned that a patchwork of state and local overlapping
laws and regulations regarding AI may lead to unintended consequences
that create uncertainty and discourage workplace innovation. SHRM
supports a uniform Federal standard that provides a clear framework for
how employers should strive to prevent unlawful bias and promote
responsible workplace applications of AI. We also support Federal
efforts to educate stakeholders about the benefits and risks of AI in
the workplace.
Conclusion
The rapid evolution of AI brings both incredible opportunities and
significant challenges. Lawmakers must work with multiple stakeholders
to identify ways to meet these challenges, and to create a regulatory
landscape that enables workers and businesses to achieve the benefits
of AI responsibly and consistent with our shared values. SHRM hopes
that lawmakers, employers, workers, and other stakeholders can work
together to achieve these goals and address these concerns.
SHRM is committed to helping employers and HR professionals
responsibly navigate this new landscape, ensuring that AI enhances the
workplace without compromising our shared values. We stand ready to
provide Congress with our expertise on workplace issues to ensure you
are fully informed about the potential consequences of forthcoming
public policy and legislative efforts. Thank you for the opportunity to
submit this testimony to the Employment and Workplace Safety
Subcommittee about this critical issue.
______
Senator Hickenlooper. Thank you, Mr. Meyer. Now, it is my
pleasure to introduce our next witness, Dr. Karin Kimbrough,
national expert on skills attainment.
She is the Chief Economist at LinkedIn. And at LinkedIn,
she is at the helm of LinkedIn's workforce research team, which
assesses trends in skills training, demand, and attainment.
Dr. Kimbrough.
STATEMENT OF KARIN KIMBROUGH, CHIEF ECONOMIST, LINKEDIN,
SUNNYVALE, CA
Dr. Kimbrough. Chairman Hickenlooper, Ranking Member Braun,
and Members of the Subcommittee, thank you for inviting me to
testify today. My name is Dr. Karin Kimbrough, and I serve as
the Chief Economist at LinkedIn.
I appreciate the opportunity to share our insights on how
AI is impacting the economy, including skills, jobs, and
industries. We believe that AI holds great promise and
potential to enhance the productivity of workers and to allow
them to focus on more meaningful aspects of their jobs.
With AI tools becoming accessible to the workforce at
large, the impact will be widespread. AI is here and it is
changing the way we work. Today, I want to discuss three key
trends from our data.
First, we are already seeing early influences of AI on jobs
and skills. Second, employers are increasingly placing a
premium on hiring and training AI talent. And third, workers
are realizing the value and benefits of acquiring AI skills.
Finally, I will touch briefly on the global AI talent
marketplace and offer ideas and policy suggestions for this
Committee to consider. At LinkedIn, we think about AI talent in
two ways, the technical builders who are developing AI tools
and the general users that will need to be AI literate.
The earliest indicators of how firms are responding to the
technological promise of AI comes from our hiring data. So far
this year, the demand for AI skills has increased. For example,
we have seen a 30 percent increase in the share of job postings
for AI technical talent compared to 2023.
The increase in demand has been even faster in the
technology industry. This increase in demand for AI technical
talent is also shifting the skills needed for today's jobs
across all industries.
Since 2015, the skills for the average job have changed by
about 25 percent, and we expect this change to accelerate. By
2030, two-thirds of what we do at work every day will look very
different. Despite this rapid change in skills, companies that
use AI are not just hiring new workers with these skills but
are also initiating internal training programs to upskill their
own workers.
One example of this employer provided training is NIQ, a
U.S. based marketing research firm. It has been using LinkedIn
learning for employee training, including AI. More than 1,900
employees participated in NIQ's AI learning challenge, logging
over 18,000 hours of AI learning content in the last year.
Additionally, employers are looking to do more training.
The Census Bureau has found that one in five companies using AI
had recently trained its staff to use it, and nearly half of
other companies plan to offer training in the next 6 months.
Workers are also taking the initiative. They are upskilling
on their own and going beyond what employers are providing as
they begin to identify the opportunities of AI in their own
work.
Over the last 6 months, we have seen a 160 percent increase
in non-technical professionals who are taking LinkedIn learning
courses to grow their AI literacy. These are workers in roles
like project manager, architect, administrative assistant, and
more than half of a million learners have enrolled in our
professional certificates for AI.
In a competitive labor market, workers see the value of
making their new AI skills known, and around the world there
are now 142 times as many members with AI literacy skills than
there were just late last year.
All of this matters on a global scale. For example, the
U.S. has a third of the world's AI technical talent and a third
of the world's workers with AI literacy skills. However, other
countries like India are developing their AI talent at a faster
rate and are catching up.
As this Committee looks ahead at policies to prepare
workers for AI, there are four areas on which I would encourage
you to focus. First, continue the bipartisan efforts to promote
policies which expand skills based hiring in this country.
Second, consider ways in which the reauthorization of WIOA
can help meet the needs of workers in the new economy. Third,
support efforts to encourage and enable employers, specifically
small businesses, to upskill their workers.
Finally, ensure that Federal labor data are able to capture
the impact of AI on our Nation's workforce accurately and in a
timely manner, including by leveraging public-private
partnerships. Thank you for this opportunity to share more
about LinkedIn's insights, and I look forward to your
questions.
[The prepared statement of Ms. Kimbrough follows.]
prepared statement of karin kimbrough
Chair Hickenlooper, Ranking Member Braun, and Members of the
Subcommittee: thank you for inviting me to testify today.
My name is Dr. Karin Kimbrough, and I am the Chief Economist at
Linkedln. I lead a team of economists and data scientists dedicated to
understanding how the world of work is evolving through what we call
the Economic Graph Research Institute (EGRI). We study workplace trends
and shifts in the macro economy, identify opportunities for
professionals in a wide range of roles and occupations, and strive to
understand the relative demand for work across multiple industries
worldwide. I have worked in the public and private sectors on
macroeconomic research and finance for over 20 years, including nearly
a decade at the Federal Reserve Bank of New York, where I was a vice
president in the Markets Group.
I appreciate the opportunity to share our insights on how the
increased deployment of artificial intelligence (Al) applications in
nearly every industry is affecting the workforce and work. Linkedln is
the world's largest professional network with more than 1 billion
members in more than 200 countries and territories worldwide. Our data
provides us with real-time, granular insights that enable us to conduct
extensive research on Al as it relates to jobs, skills, workers, and
the emerging global Al economy.
We believe Al holds great promise and potential to enhance the
productivity of workers, and to allow workers to focus on more
meaningful aspects of their jobs.
There are three broad points which I would like to highlight from
our insights and other research, which are relevant to today's hearing.
First, our early observations suggest Al is already impacting jobs
and skills; second, employers are placing a premium on Al talent and
employer-driven Al training; and third, workers are increasingly
realizing the value and benefits of acquiring Al skills.
Finally, I will touch upon the global context of Al supply and
demand, and the policies we encourage this Committee to consider to
further expand access and opportunities for Al skilling.
I. Al is Impacting Jobs and Skills
Almost a year ago, this Subcommittee held a hearing focused on Al
and the future of work, with the expectation that Al, and Generative Al
in particular, would have a growing impact on jobs and workers. Over
the past year, we have seen this become a reality.
Al is here, and it is changing the way we work. The earliest
indicators of this evolution come from our data around the hiring of Al
technical talent--meaning those with Al engineering skills, who are
critical to enabling companies to build the right tools for Al
implementation and adoption at scale.
LinkedIn hiring data show that there was a 6 percent increase in
hiring for Al technical talent in the United States in the first 12
months after ChatGPT launched. This reflects an acceleration in demand
for Al expertise and is an important initial step in firms' success in
deploying Al more broadly. \1\ This hiring increase coincided with a
hiring slowdown in the Technology industry at large (^10 percent y/y),
and the United States overall (^10 percent y/y). \2\ Further, data over
the last 8 years show that hiring for Al technical talent has increased
by more than 300 percent globally. \3\
---------------------------------------------------------------------------
\1\ Linkedln Economic Graph Research and Insights. See also
Stanford University, Artificial Intelligence Index Report 2024, https:/
/aiindex.stanford.edu/wp-content/uploads/2024/05/HAI--Al-lndex-Report-
2024.pdf, and OECD, OECD.AI Policy Observatory, https://oecd.ai/en/
data'selectedArea=ai-jobs-and-skills&selectedVisualization=ai-talent-
concentration-by-country.
\2\ Id.
\3\ Microsoft and Linkedln, 2024 Work Trend Index Annual Report,
May 8, 2024, https://www.microsoft.com/en-us/worklab/work-trend-index/
ai-at-work-is-here-now-comes-the-hard-part.
Beyond demand for Al technical talent, we see an emerging impact of
Generative Al and Al tools on the broader workforce. In particular, the
wider accessibility of these new technologies to a non-technical
workforce using equipped smartphones and laptops suggests that the
speed of diffusion will be faster than prior technological innovations.
In other words, the impacts of this technology will be widespread,
---------------------------------------------------------------------------
crossing industries, jobs, and skills.
The extent to which Al has penetrated the global workforce is quite
remarkable. Microsoft and Linkedln released the 2024 Work Trend Index
Report, which surveyed 31,000 people across 31 countries, and found
that 75 percent of global knowledge workers report using Generative Al
at work. Equally striking is the speed at which this is happening. The
survey found that 46 percent of those Al users report that they started
using it less than 6 months ago. \4\
---------------------------------------------------------------------------
\4\ Microsoft and Linkedln, 2024.
As Al is becoming more ingrained in our daily work, we're seeing
new jobs being created and the skill sets for roles shift. The number
of companies with a ``Head of Al'' position has tripled in the past 5
years and grew by more than 28 percent in 2023. Globally, 12 percent of
recruiters now say they are creating new roles tied specifically to the
use of Generative Al. \5\ As an example of how jobs are changing, our
data show the skills needed for the average job have changed by 25
percent since 2015. By 2030, we expect skills to change by more than
two-thirds, with tools like Generative Al accelerating this change. \6\
In other words, for many of us, two-thirds of what we do at work every
day will look very different in just a few short years.
---------------------------------------------------------------------------
\5\ Id.
\6\ Id.
We also expect Al and Generative Al to impact a wide number of the
jobs and occupations held by workers today. To better understand what
this impact is likely to look like, we have analyzed data across our
platform to estimate the percentage of occupational skills that are
complemented by Al and the percentage of skills that are replicable by
Al. What we found, as highlighted in the chart below, is the impact Al
will have on jobs will vary considerably. Some jobs, like translator,
are likely to experience disruption because a significant portion of
the skills can be replicated by Generative Al. As a result, these
workers will need extensive training to develop new Al-related skills
to effectively integrate Al into their workflows. On the other hand,
workers in other occupations, like physical therapists, will be largely
insulated from Generative Al because many of the skills they use are
complemented by Al, but not replicated by it. However, over time, we do
expect that the vast majority of occupations--even ones that appear
insulated for now--will eventually incorporate Generative Al into their
tasks, with some occupations transforming more than others. \7\
---------------------------------------------------------------------------
\7\ Karin Kimbrough and Mar Carpanelli, Preparing the Workforce
for Generative Al, August 23, 2023, https://economicgraph.linkedin.com/
content/dam/me/economicgraph/en-us/PDF/preparing-the-workforce-for-
generative-ai.pdf.
[GRAPHIC] [TIFF OMITTED] T7256.100
By extension, this also means the workforces of certain
industries are likely to be more directly impacted by Al than others.
For example, we found that 50 percent of workers in the Financial
Services industry are likely to face Al disruption in their jobs,
compared to 36 percent of workers in the Technology industry and just
18 percent of workers in Accommodation and Food Services. \8\ This
points to the need for and value of working within sectors not only to
better understand the potential impacts of Al, but to address worker
needs, which I will discuss further below.
---------------------------------------------------------------------------
\8\ Kimbrough and Carpanelli, 2023.
Despite these projected disruptions, to date, recent Federal data
suggest that the augmentation and disruption of these skills is not
necessarily leading to people losing their jobs. According to recent
U.S. Census Bureau data, employers in the U.S. have few plans to reduce
their total employment due to Al. In fact, the percentage of employers
anticipating an increase (6.5 percent) in employment edges out those
anticipating a decrease (6.1 percent). \9\
---------------------------------------------------------------------------
\9\ U.S. Census Bureau, Business Trends and Outlook Survey, 2024,
www.census.gov/programs-surveys/btos.html.
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]
Consistent with these findings, our research shows that the
adoption by firms of GitHub Copilot, a Generative Al powered code
completion and automatic programming tool, boosts hiring by about 3
percentage points, particularly for entry-and senior-level engineers,
with firms hiring more software engineers overall. It also leads to a 7
percent increase in job postings for software engineers on Linkedln,
without displacing existing workers. \10\
---------------------------------------------------------------------------
\10\ Matthew Baird, Mar Carpanelli, Brian Xu, and Kevin Xu, GitHub
Copilot and the Future of Work: A Working Paper, September 2023,
https://economicgraph.linkedin.com/content/dam/me/economicgraph/en-us/
PDF/github-copilot-working-paper.pdf.
II. Employers are Placing a Premium on Al Talent and are Focusing on
Training
Employers are starting to see the advantage of Al and its capacity
to increase productivity and, therefore, the need to hire workers with
technical and broader Al literacy skills, by which we mean the
familiarity with and ability to use Al tools. The 2024 Work Trend Index
found that 66 percent of business leaders with decision-making
authority said they wouldn't hire someone without Al skills, and 71
percent said they'd rather hire a less experienced candidate with Al
skills than a more experienced candidate without them. \11\
---------------------------------------------------------------------------
\11\ Microsoft and Linkedln, 2024.
If these global survey results hold true in the coming years, we
expect to see the demand for those with Al skills, both in engineering
and literacy, to quickly outpace supply given the labor pool largely
---------------------------------------------------------------------------
lacks these skills.
In the face of this challenge, a growing number of employers are
realizing the need not only to seek out new workers with Al skills, but
more importantly, to develop their own Al workforce--including both Al
technical and Al literate talent.
As an example, NIQ, a U.S.-based marketing research firm out of
Chicago that uses Linkedln Learning, has been working to train their
workers on Al. In the last 12 months, NIQ employees have watched more
than 73,000 total hours of Linkedln Learning content and, when looking
at the types of courses NIQ employees are taking, five of the top ten
courses are on Al. Additionally, more than 1,900 employees registered
for NIQ's Al Learning challenge--during which employees watched more
than 18,000 hours of Al learning content.
Recent data from the U.S. Census Bureau point more broadly to this
trend. The Business Outlook and Trends Survey found that the percentage
of companies reporting plans to hire staff already trained in Al in the
next 6 months will quadruple compared to those who reported hiring
trained staff in the past 6 months. While this change is significant,
this trend is dwarfed by the percentage of companies who report
planning to offer Al training to their workers, which will double from
21 percent reporting having trained current staff in the past 6 months
to 42 percent planning to provide training in the next 6 months. This
tells us that companies know they need to train their own workers in Al
because outside talent is still skilling up and not yet available in
large numbers. \12\
---------------------------------------------------------------------------
\12\ U.S. Census Bureau, 2024.
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]
Source: U.S. Census Bureau, Business Outlook Trends Survey \13\
---------------------------------------------------------------------------
\13\ Id.
While this data points to a positive trend, the fact remains that
employers in the U.S. could be doing more to ensure their employees
have the Al skills necessary to fully adopt Al into their work. In
fact, we found that only a quarter of Learning and Development teams
globally plan to offer training programs on how to use Generative Al
this year. \14\
---------------------------------------------------------------------------
\14\ Linkedln Learning, Linkedln Workplace Learning Report 2024,
www.learning.linkedin.com/content/dam/me/business/en-us/amp/learning-
solutions/images/wlr-2024/Linkedln-Workplace-Learning-Report-2024.pdf.
Business leaders who fail to devote the resources necessary to
reskill and upskill their Al talent risk falling behind companies who
are making these investments and seeing productivity gains from
leveraging this technology.
Ill. Workers are Increasingly Seeing the Value of Acquiring Al Literacy
Skills
Our data show that workers recognize the value of Al skills in an
increasingly competitive labor market. We observe this keen interest by
workers to upskill firsthand in our own Linkedln Learning data.
Linkedln Learning currently offers over 1,000 courses on Al, from
learning basic Al principles, concepts, and applications to advancing
skills in machine learning. Since January 1, 2023, the number of people
taking Linkedln Learning courses on Al has grown by 13 times. \15\ As
of earlier this year, the use of Linkedln Learning courses designed to
build Al literacy skills spiked 160 percent among non-technical
professionals, with workers in roles like project managers, architects,
and administrative assistants looking to skill up the most. \16\
---------------------------------------------------------------------------
\15\ Linkedln Learning, 2024.
\16\ Microsoft and Linkedln, 2024.
Workers recognize that acquiring Generative Al and other Al skills
not only makes them more competitive but also can render work more
productive, safer, and potentially more interesting. According to the
Work Trend Index, workers who are using Al tools report that Al helps
them save time (90 percent), focus on their most important work (85
percent), be more creative (84 percent), and enjoy their work more (83
percent). \17\
---------------------------------------------------------------------------
\17\ Microsoft and Linkedln, 2024.
Workers also understand the value of making it known to their
current or prospective employers that they have acquired these
competencies. Our data show there are now 142 times as many members
with Al literacy skills on their Linkedln profiles than there were last
year. \18\
---------------------------------------------------------------------------
\18\ Karin Kimbrough, Al at Work Has Arrived, May 8, 2024, https:/
/economicgraph.linkedin.com/blog/ai-at-work-has-arrived.
While it is encouraging that so many workers are able to seek out
their own training opportunities, we know this is not enough. There are
still barriers in place that prevent access to these opportunities
across multiple occupations. We believe a more coordinated and
widespread effort is necessary to ensure that all workers have the
ability to gain the necessary Al skills.
The Global Context
With respect to Al skills globally, we see that the U.S. has 31
percent of the world's Al technical talent and that other countries are
picking up the pace. For instance, the concentration of this talent in
the U.S. increased by 13 percent from 2022 to 2023. Over the same
period, the growth in Al technical talent was faster in the United Arab
Emirates (29 percent y/y), India (24 percent y/y), Saudi Arabia (21
percent y/y), the United Kingdom (16 percent y/y), and Canada (15.4
percent y/y). The pace of growth of Al technical talent was relatively
modest in Germany (9.5 percent y/y) and France (2.1 percent y/y). \19\
---------------------------------------------------------------------------
\19\ Linkedln Economic Graph Research and Insights. See also
Stanford University and OECD.
Similarly for members with Al literacy skills, over 30 percent of
members with these skills are in the United States, closely followed by
India at 29 percent. \20\ We expect demand for this talent to continue
to grow in the U.S. and around the world.
---------------------------------------------------------------------------
\20\ Id.
---------------------------------------------------------------------------
Policy Considerations
I commend this Subcommittee for seeking to better understand the
impact and opportunities of Al. As you continue your work in this area,
we would encourage you to consider policies that accelerate Al adoption
across our Nation, and ensure it benefits those who don't have access
to training or who need additional support to gain these skills.
The Committee's current effort to reauthorize the Workforce
Innovation and Opportunity Act (WIOA) provides one way in which to
support such individuals from across all sectors, as the economy shifts
toward an Al-driven future.
Congress can also promote the adoption of skills-based hiring as a
way to provide more opportunities to those who may have in-demand Al
skills but may lack a traditional 2-or 4-year degree. Skills-first
hiring allows companies to tap into a broader talent pool, including
individuals who have gained relevant skills through non-traditional
pathways such as bootcamps, online courses, or self-study. This
approach can be particularly beneficial in rapidly evolving fields like
Al, where the demand for specific skills often outpaces the supply of
formally qualified candidates.
We appreciate the opportunity to work with Members of the
Subcommittee on this and other legislation and look forward to
collaborating on support for workers and employers who seek to take
advantage of the benefits Al has to offer, including small businesses,
many of which may benefit from the potential for Al tools to increase
efficiencies and productivity.
Finally, we would encourage you to consider how to bolster Federal
data, including through public-private partnerships, to better
understand how Al is impacting the workforce.
Thank you for the opportunity to join you here today and I look
forward to answering any questions.
______
Senator Hickenlooper. Thank you, Dr. Kimbrough. Now we will
hear from Mr. Kotran. Mr. Kotran serves as the Chief Executive
Officer for aiE-D-U. I guess you could say aiEDU, an education
based nonprofit that partners with education and technology
experts across the country to provide AI literacy training for
our educators.
Mr. Kotran.
STATEMENT OF ALEX KOTRAN, CHIEF EXECUTIVE OFFICER, AIEDU, SAN
FRANCISCO, CA
Mr. Kotran. Thank you. Good afternoon, Members of the
Subcommittee. I want to--really appreciate the opportunity to
testify today. My name is Alex Kotran and CEO of the AI
Education Project.
We are a nonprofit and our mission is to bring AI readiness
to classrooms nationwide by empowering our teacher workforce.
Today, I appear before you today to underscore the urgent need
to massively expand investment in teacher training and
professional development.
We need to ensure that we are preparing students for the
profound changes that are coming with the disruption that AI is
going to bring to our society and indeed our workforce. aiEDU
is a 501(c)(3) nonprofit.
We were founded 5 years ago, and we have been leading a
national movement to advance AI readiness and AI literacy. We
define AI literacy as the collection of skills and knowledge
that a person needs to confidently understand, ethically use,
and critically evaluate artificial intelligence in a world in
which AI is becoming more and more ubiquitous.
We define AI readiness as the durable skills and the
competencies that students, teachers, and school systems are
going to need in order to apply AI literacy. Our work is
anchored on a sophisticated national grassroots operation, and
we partner with some of the largest districts and education
systems in the country.
We stand at the dawn of a transformative era. The AI
revolution is going to define the trajectory of an entire
generation of students and workers, and it is going to
fundamentally change the skills they need to compete in that
workforce.
The urgency and the pace of this challenge is hard to
overstate. Less than 2 years ago, large language models or LLMs
were virtually nonexistent. And today, hundreds of millions of
people are using them every day, and we are starting to see
meaningful applications that are resulting in significant
impacts on the future of work.
I am going to highlight one example. Last month, Amazon CEO
Andy Jassy shared that their AI coding assistant, Amazon Q,
made eye popping advances in replacing and augmenting the
company's software engineers. They replaced the equivalent of
50 developer days for a very simple task uploading a Java 17
application.
In total, Jassy predicted that the tool has saved the
equivalent of 4,500 developer years and more over 80 percent of
the code that it writes does not require any editing by humans.
Research into the potential impacts of LLMs in the workforce
suggests wide ranging applications and displacement,
concentrated in some of the most common knowledge work careers
in the U.S..
I think most concerning, the National Bureau of Economic
Research has conducted a study that found that historically the
pace of automation driven job displacement, it doesn't happen
gradually, but rather it happens in these acute bursts in the
months following recessions.
I fear that the next recession, whenever it comes, will be
the moment that we must all truly face the music. So what do we
do? Our organization has learned in more than 5 years of
working with educators and school leaders across the country
that there is no one size fits all approach.
We need to work state by state and in many cases district
by district to create models that scale and massively invest in
our teacher workforce.
We have also leveraged partnerships with state agencies,
school districts, and community organizations, all with the
goal of building capacity amongst the teachers, the
administrators, and the systems that support them to build the
capacity and the competencies that they need to actually
deliver and build the skills that ultimately students and
workers are going to need.
Our research underscores the urgency. What we found is that
72 percent of teachers when surveyed report only being
moderately comfortable with AI at best. For our Country to
build a base of new skills, we need to significantly increase--
decrease that number, rather. The goal of our education system
must be to include an expanded focus on durable skills.
We have to focus on critical thinking, communication,
collaboration, resilience, adaptability, and digital
citizenship, and these are best developed through live,
engaging classroom activities facilitated by real educators. We
believe that a range of AI competencies must be widely
understood and achieved in our education system.
These include knowing the basics of AI, yes, but also being
able to understand when to use AI, and when to bring in and
leverage the human advantage.
In terms of recommendations, we really need to push forward
a coordinated and concerted effort at the national, state,
regional, and local levels to elevate the conversation and push
for action around the skills and training for both teachers and
students, and indeed our workers.
Congress can create the right conditions first and foremost
by routing more funds and creating the supports and the
necessary investments, in particular to provide schools with
resources for teacher training and professional development.
Funding aside, the hyperlocal nature of this challenge
demands a national movement to elevate the conversation about
AI readiness, and I believe all of you sitting here today have
the trust and influence at your disposal to champion this
message, both here in Washington and at home.
I want to close by underscoring that while educators are
understandably concerned about the implications of AI on their
jobs, at aiEDU, we firmly believe that teachers are critical to
helping students build the skills they need to succeed.
Put simply, our teachers are at the very heart of any
solution to the future of work. Thank you, and I look forward
to your questions.
[The prepared statement of Mr. Kotran follows.]
prepared statement of alex kotran
Good morning, Chairman Hickenlooper, Ranking Member Braun, and
distinguished Members of the Committee. Thank you for the opportunity
to testify today on the critical topic of artificial intelligence in
education.
I want to first applaud the Committee for its recent work advancing
two bills--the Teachers LEAD Act of 2023 and the STEM Education in
Accounting Act--which demonstrate a deep understanding, which I share,
about the vital roles that teacher leadership and career exploration
play in preparing students for the future. By supporting these
initiatives, the Committee is taking important steps to enhance
education quality and expand career opportunities, reflecting a strong
commitment to strengthening our education system and workforce
readiness.
My name is Alex Kotran, and I am the CEO and founder of aiEDU, an
organization dedicated to bringing AI expertise to classrooms
nationwide. I have been deeply involved in the AI space since 2015. I
co-founded aiEDU in 2019, well before the public introduction of large
language models (LLMs). Our vision is to ensure that every student is
ready for the age of artificial intelligence.
Today, I appear before you to underscore the urgent need to
massively expand investment in teacher training and professional
development to ensure that our schools and educators have the capacity
to prepare students for the profound changes and disruption that AI
will bring to our society and economy.
The Reach and Impact of aiEDU
I'd like to begin this testimony with a brief summary of aiEDU's
work, as our experience is highly relevant to the topic being
discussed. Founded in 2019, The AI Education Project is a national
501c3 nonprofit organization that has been leading a national movement
to advance AI Readiness and AI Literacy for all students through
concerted efforts to build capacity with teachers and education
systems.
We define AI Literacy as the collection of skills and knowledge
that a person needs to confidently understand, ethically use, and
critically evaluate artificial intelligence in a world in which AI is
becoming more and more ubiquitous.
We define AI Readiness as being when students, teachers, and school
systems have the ability and underlying skills to apply AI Literacy to
professional and personal endeavors. A person is AI ready when they
understand the interdisciplinary impacts of AI and how to apply their
human advantage alongside evolving technology, to leverage
collaboration, creativity, and self-advocacy alongside AI to achieve
their life and career potential.
At aiEDU we've built a sophisticated grassroots operation that
reaches tens of thousands of educators, anchored in partnerships with
some of the largest districts and education systems in the country.
This includes: Prince George's County Public Schools, the State of Ohio
(via a coalition that includes the Lt. Governor's office and Dept. of
Education), the State of Colorado (via the Colorado Education
Initiative), and the National Rural Education Association (which
reaches 9M students). Collectively, our reach is now impacting hundreds
of thousands of students and tens of thousands of teachers in every
single state in the U.S.
aiEDU has built a coalition of early adopters for AI Literacy in K-
12 that are now leading the country with demonstrated models of tiered
support enabling delivery of AI curriculum and professional
development. We make an impact mainly through curriculum
implementations and teacher trainings that directly impact students. We
have focused on the K-12 space because it is possible to create
equitable reach for all students via established systems, and is a
natural middle ground between government, nonprofits, and industry.
And, of course, because student success is inherently non-partisan.
We've developed a template in which aiEDU provides critical support
to empower districts and non-profit groups: We supply the content,
subject matter expertise, and training that empowers our partners to
scale delivery of AI expertise and instructional support across their
networks. Our focus on districts and large systems is also fundamental
to our mission to reach all students, as we intentionally select
regions with large populations of students who are at risk of otherwise
being left behind.
Our goal is to bring AI expertise to classrooms nationwide.
Teachers and education leaders trust aiEDU because we don't build or
sell AI products (of which there are now thousands). Instead, we help
leaders close a critical gap that every major district or K-12 system
in the country faces: Preparing teachers and students for the age of
AI.
Over the past year alone, our organization has made significant
strides in promoting AI education. We've conducted more than 120
events, including summits, workshops, and seminars, reaching more than
10,000 educators and administrators. Our aiEDU tools and curriculum
have been downloaded by over 1,350 educators, impacting 231,000
students across all 50 states. We've launched major district programs
in New York City and Prince George's County, expanded statewide
initiatives in Ohio, Hawaii, and Colorado, and partnered with
organizations serving rural and indigenous communities.
The Critical Importance of AI Readiness
We stand at the dawn of a transformative era. The AI revolution
will define the trajectory of an entire generation of students and
fundamentally change the skills they need to compete in the workforce.
The urgency of this challenge is hard to overstate.
Early indications about the capabilities of large language models
are startling and suggest that we are on the cusp of massive disruption
to our workforce and social fabric in the coming decade. The pace of
development is rapid, with a steady drumbeat of new models featuring
increasingly massive training dataset along with powerful new AI
graphical processing units (GPUs) that reflect the hundreds of billions
of dollars of investment that has been pouring into the space.
Earlier this month, OpenAI revealed its ``o1'' model
family, which are designed to employ complex reasoning through
a ``chain-of-thought'' process that, in short, has enabled
their newest LLM to exceed PhD-level accuracy on benchmarks of
physics, biology, and chemistry problems, as well as beat the
vast majority of humans at competitive programming and math
competitions.
Less than 2 years ago, Generative AI was virtually non-existent for
consumers. Today, it is widely accessible, rapidly evolving, and
starting to demonstrate meaningful applications that could result in
significant impacts on the future of work:
Software Engineering: Last month Amazon CEO Andy Jassy shared that
their AI assistant, Amazon Q, made eye-popping advances in replacing
and augmenting the company's software engineers, with average time to
upgrade an application to Java 17 dropping from 50 developer days to
just a few hours. Jassy stated that this efficiency saved the company
the equivalent of 4,500 years of developer work, and that nearly 80
percent of AI-generated code reviews were finalized without any
additional changes.
Healthcare: There are myriad studies demonstrating LLM's ability to
provide accurate diagnoses. A recent study by Google Research in
collaboration with Google DeepMind revealed that an LLM developed with
conversational and collaborative capabilities was able to improve
diagnostic reasoning and accuracy by clinicians working on complex
medical situations, with the performance of clinicians augmented with
the LLM achieving 59.1 percent accuracy, compared with 33.6 percent of
clinicians working unassisted.
Entertainment: The rapid advances in generative AI and diffusion
models, which enable text-to-image, text-to-speech, and text-to-video
content synthesis, have rocked Hollywood and the entertainment
industry, with a recent survey of 300 executives reporting that 75
percent indicated AI tools will support the elimination, reduction, or
consolidation of jobs at their companies. That same report estimates
that 204,000 entertainment jobs will be affected by AI within the next
3 years alone.
Anecdotes aside, research into the potential impact of LLMs on the
workforce suggest wide-ranging applications and displacement,
concentrated in some of the most common knowledge work careers in the
U.S.
A study by OpenAI and the University of Pennsylvania investigating
the labor market impacts of LLMs identified a number of occupations
with significant exposure to AI-enabled automation of tasks, including:
Financial Analysts, Writers and Authors, Web and Digital Interface
Designers, Accountants and Auditors, Engineers, Clinical Data Managers,
Legal Secretaries, Administrative Assistants, to name a few.
Pew estimates that 19 percent of American workers are currently
employed in jobs that are ``the most exposed to AI, in which most
activities may be either replaced or assisted by AI.'' Researchers
noted that ``jobs with a high level of exposure to AI tend to be in
higher-paying fields where a college education and analytical skills
can be a plus.''
In the years ahead, industries that employ the majority of
Americans will improve their AI implementations to increase
productivity and efficiency. This will change the skills workers need
in order to keep up, compete, and thrive. Workers will increasingly
have help from AI assistants, powerful new tools, and back-end process
automation that displaces routine tasks. Historically, automation
doesn't happen gradually; since 1980, the vast majority of job loss in
occupations hit by automation took place immediately following
recessions. Given it has been 16 years since our last serious
recession, we should anticipate acute impacts from AI on employment
whenever we experience our next one.
AI Readiness is important beyond preparing students to succeed in a
changing world of work--AI is already having a profound, if under-
appreciated, impact on our kids' social lives in ways we might not have
anticipated.
Character.ai is now the third most popular AI tool, with millions
of people spending multiple hours per day interacting with AI chatbots.
This trend is deeply concerning at a time when we are dealing with an
epidemic of loneliness, isolation, and depression among students. And
we are yet to see the impacts that new voice-enabled products will have
amid warnings from companies like OpenAI that anthropomorphism of AI
could ``reduce [users] need for for human interaction--potentially
benefiting lonely individuals but possibly affecting healthy
relationships.''
Natasha Singer at the New York Times has reported on the even more
disturbing phenomenon of schools encountering instances of sexual abuse
conducted via nudification apps, which enable teens to ``pervert real,
identifiable photos of their clothed female classmates, shown attending
events like school proms, into graphic, convincing-looking images of
the girls with exposed A.I.-generated breasts and genitalia.''
Michelle Culver, the founder of The Rithim Project, a nonprofit
working to advance human connection for young people, recently outlined
the table stakes: ``In 2021, 22 percent of high schoolers reported that
they had considered suicide; 4 in 10 said they experienced persistent
feelings of sadness or hopelessness. . . only 27 percent of American
men have at least six close friends; 30 years ago, that figure was more
than twice as high. Some 15 percent report having none at all. Across
the board, the trend lines are going in the wrong direction.''
Preventing students from eschewing human relationships in favor of
AI chatbots that will never reject them or make them feel awkward is
likely as difficult as banning social media or gaming, which is to say,
impossible. Likewise, deepfake tools are often run by shell companies
with servers in opaque jurisdictions overseas that are difficult or
impossible to block. Educators are therefore a critical component of
any strategy to build emotional resilience, social skills, and
knowledge that students will need when navigating the increasingly
complex digital world.
Given these rapid changes, AI Literacy will be critical to building
resilience and awareness among students. We've seen the impact of
social media on society, and we must learn from those experiences as we
navigate the AI revolution.
Building Capacity for AI Readiness Centers on Teacher Training and
Professional Development
What our organization has learned in over 5 years of working with
students across the country is that there is no one-size-fits-all
approach to scaling AI readiness. However, building increased system
capacity is key for all schools.
We need to work state by state and, in some cases, district by
district, to create models that scale, ensuring all students, teachers,
and school systems have access to quality AI curriculum and are
prepared for what's ahead.
In Ohio, for example, aiEDU has built a partnership with the
Department of Education and Workforce, the Governor's office, and
hundreds of schools across the state in collaboration with educational
service centers. This approach has allowed us to raise system capacity
across the state in a way that makes sense for Ohio's specific needs.
Given the stakes, it is also all the more critical that we work to
build capacity in even the most hard-to-reach communities.
aiEDU has leveraged partnerships with community-based organizations
that work across state lines, including Indigitize, which is partnering
with our organization to build capacity with indigenous students, and
the National Rural Education Association to reach students in rural
settings.
At the center of our approach to build capacity, and reach all
students is teacher training and professional development. It is clear
from our efforts and from our quantitative and qualitative research
that teachers and administrators want--and need--support and access to
professional development on AI.
While large language models are still maturing and may not yet be
suitable for scaled implementation at the district level, systems can
start by focusing on building AI literacy. This lays critical
foundational groundwork that is essential to any successful
implementation of AI tools or the policies that enable their safe and
responsible use.
Our own research underscores the urgency of this need. The current
and future workforce implications are significant given that about 13
percent of the American workforce works in public elementary and
secondary school. In a survey of thousands of teachers, we found that
an overwhelming majority--more than 72 percent--said they were only
moderately comfortable with AI tools. For our Country to build the base
of new skills we need and to prepare our students, teachers and
administrators adequately for what is ahead, this number must change
dramatically.
Focusing on Durable Skills
The goal of our education system must include an expanded focus on
durable skills that will be critical to enabling people to augment AI
tools that are certain to become increasingly commonplace in the
workplace. These skills--critical thinking, communication,
collaboration, computational thinking, resilience, adaptability, and
digital citizenship--are best developed through live, engaging
classroom activities facilitated by educators.
In order to cultivate AI Literacy and build AI Readiness that will
serve a generation of workers defined by AI technology, we believe that
a range of AI competencies must be widely understood and achieved in
our education system.
First, students must know the basics of AI--knowing what AI is, how
to use it, understanding how AI works along with its foundations in
computing, and perhaps most important to be continuously curious about
AI.
Second, when they know the basics, students must be able to think
critically with AI--spotting and questioning bias, examining AI's use,
and being continuously curious about AI's influence their lives,
careers, and society. In practice, that will mean things like
identifying why outputs from AI tools have discrepancies (e.g., values
and bias in creating data sets), critiquing AI systems for embedded
biases and propose ways to make them more inclusive and ethical, and
critically assessing AI outputs, considering potential biases and
limitations.
Thirdly, a critical component for students and the workforce in an
era defined by new technology to know and truly understand your human
advantage. We must ensure that students are innovative and persistent
in their pursuit of solutions, are able to synthesize interdisciplinary
and AI knowledge to solve problems in a holistic way, and to leverage
their own ability to collaborate and self advocate when they navigate
their careers.
There are opportunities to connect durable skills and AI Readiness
together--and it's important to make these connections accessible and
to motivate educators to work with them. A growing number of teachers
in our research--54 percent--recognize AI literacy as an important or
essential skill for their students future. But fewer do so than other
important life skills for students such as social emotional skills (92
percent) critical thinking and problem solving (95 percent) career
exploration (88 percent) and digital citizenship and media literacy (75
percent).
Recommendations
We urgently need coordinated and concerted efforts at the national,
state, regional, and local levels to elevate conversation and action
around skills and training for both teachers and students. While there
isn't a clear policy or approach at the Federal level alone which could
address this challenge, Congress can nonetheless create the conditions
by elevating the conversation and routing more funds and support for
states and districts to make the necessary investments.
We support several key policy changes:
1. Establishment of Federal AI standards to provide common
benchmarks for AI literacy and implementation in education.
2. Development of aligned data standards to enable centralized
tracking of progress in AI education.
3. Increased investment and support for upskilling educators in
AI technologies and pedagogies.
It's crucial to note that while educators are understandably
concerned about the implications of AI on their jobs, we firmly believe
that teachers are critical to helping students build the skills
required to succeed in an AI-driven world. Teachers must never be
displaced by automated tools. This stands in contrast to the rhetoric
coming from some in the tech community who envision a world where our
education system is centered on ubiquitous AI tutors.
______
Senator Hickenlooper. Thank you, Mr. Kotran. Finally, we
have Mr. Denzel Wilson. Mr. Wilson serves as the Grassroots
Community Manager for Seed AI, where he leads efforts to
facilitate community based AI literacy training through a
coalition of corporate and higher education partners. Mr.
Wilson.
STATEMENT OF DENZEL WILSON, GRASSROOTS PROGRAM MANAGER, SEED
AI, WASHINGTON, DC
Mr. Wilson. Thank you, Chairman Hickenlooper, Ranking
Member Braun, and Members of the Subcommittee. Thank you for
inviting me to testify today. My name is Denzel Wilson, and I
am the Grassroots Program Manager for Seed AI.
I am grateful for the opportunity to speak on this
critically important topic of preparing the American workforce
for AI. Seed AI is a nonpartisan, nonprofit organization
working at the forefront of AI policy and governance.
We work with policymakers, scientists, academics, and the
private sector to design policies, programs, and tools to
ensure AI benefits all Americans. Today's topic is a personal
one for me because I grew up in Kankakee, Illinois, a county
right outside of Chicago, where opportunities were scarce, and
career options were limited.
After graduating high school, I faced several obstacles,
including being laid off during the COVID-19 pandemic. That led
me on an unexpected path. This setback sparked a pivotal career
change. So in the fall of 2020, I decided to pursue my interest
in AI by joining the AI program at Houston Community College.
My efforts culminated in becoming the first graduate of HCC
associate's in Applied AI degree, a first not just for the
college but for the entire State of Texas. This experience
paved the way for my current role in Seed AI, which I joined in
2023.
Our goal with grassroots work at Seed AI is to turn anxiety
and intrigue and turn that intrigue into opportunity. This work
has revealed key lessons that can help prepare the American
workforce for AI.
First, students from community colleges, Historically Black
Colleges and Universities, and other underrepresented
demographics are hungry and capable of harnessing the power of
AI as their advantage. And second, when presented with new
perspectives and opportunities, students from these
demographics thrive in environments that require
resourcefulness and creativity.
We do this in several ways, including running our Hack the
Future Initiative to inspire and educate participants about the
potential of AI. These events have collectively reached
hundreds of students, providing them with hands on experience
and valuable insights to the world of technology and
innovation.
Seed AI also helped conceive and organize the largest ever
generative AI red team at the Hack Convention DEFCON in 2023.
This featured over 2,000 participants, including 220 students
and entrepreneurs from Tulsa, HBCUs, and community colleges
flown in by Seed AI and our partners.
This unprecedented event demonstrated that there is a
pathway for people from all walks of life to participate in
shaping AI. Seed AI has also implemented a robust micro grant
program to fill gaps we saw in the AI skill landscape.
By the end of 2024, Seed AI will distribute over 90,000
micro grants across nine community colleges and universities,
directly investing in the future of AI education, and ensuring
that a diverse range of students can participate in and
contribute to the AI revolution. Lowering the barriers to
acquiring AI skills is just one piece of the puzzle.
Skills alone will only go so far. We need to invest in
creating the infrastructure that all who wish to can access
computational and data reader sources necessary to meaningfully
engage and co-create with AI.
We believe community colleges are uniquely equipped to
create opportunities for people that are capable of succeeding
as scientists and engineers, but do not have access to pathways
that lead into those careers. What the NSF calls the missing
millions.
This is why Seed AI is proud to partner with the NSF funded
National Applied Artificial Intelligence Consortium, an
initiative launched by three community colleges, all of which
were awarded micro grants from Seed AI in 2024.
Applied AI skills are the key to unlocking the benefits of
fundamental AI advancements, as well as well-paying jobs and
sustainable careers for all Americans. Seed AI is also proud to
participate in the Tulsa hub for Equitable and Trustworthy
Autonomy, which was recently awarded $51 million by the U.S.
Economic Development Administration.
We view partnerships like the Tulsa Hub and the National
Applied AI Consortium as critical last mile components of
things like the National AI Research Resource. These
partnerships address the critical need to ensure that the
research from the NAIR can be co-developed and leveraged by
communities across the country.
There is both incredible appetite and untapped potential in
American workers and students nationwide to participate in and
benefit from AI. Seed AI is eager to work with Members of this
Subcommittee to help realize the vision of preparing the
American workforce for AI and leverage the lessons we have
learned from our grassroots work.
There are several key opportunities we think would help
make this happen as fast and as equitably as possible that I
would be happy to discuss today. Most importantly, we encourage
the Subcommittee to leverage its power of attention.
Progress on this topic requires robust public-private
partnerships and tapping all stakeholders in the Government,
industry, and academia to make changes to facilitate greater
participation in and access to AI opportunities.
We can transform the AI ecosystem to one that is more
inclusive and participatory, more geographically diverse, and
more beneficial for all Americans. Thank you to Chairman
Hickenlooper, Ranking Member Braun, and Members of the
Subcommittee. I look forward to answering your questions.
[The prepared statement of Mr. Wilson follows.]
prepared statement of denzel wilson
Chairman Hickenlooper and Ranking Member Braun.
Thank you for inviting me to testify today before the Subcommittee.
My name is Dezel Wilson, and I am the Grassroots Program Manager for
SeedAI. I am grateful for the opportunity to speak on this critically
important topic of preparing the American workforce for AI.
SeedAI is a non-partisan, non-profit organization that works at the
forefront of artificial intelligence policy and governance,
collaborating with policymakers, scientists, academics, and the private
sector to design policies, programs, and tools to ensure AI benefits
all Americans.
Ensuring that AI benefits Americans does not just mean that
Americans can access the latest technologies. It means giving Americans
the tools to participate in the development of AI, to harness the
potential of AI for their empowerment, to democratize access to the
skills necessary to make AI work for them, and to expand economic
opportunities for new students and existing members of the workforce
alike.
This topic is a personal one for me. I grew up in Kankakee, IL, a
county outside of Chicago, where opportunities were scarce and career
options were limited. Outside of football, technology was my passion
growing up. Over the years, I witnessed firsthand how the steady
evolution of technology was transforming the world around me and how
the rise of AI has created so many unanswered questions in our world
today. I am grateful to be a part of what some call the ``Fourth
Industrial Revolution,'' working to tap into our human instinct for
innovation to ensure we create the necessary infrastructure,
partnerships, and policies so that AI serves the needs of the many, not
just the few.
I'd like to highlight three things in my testimony today:
1. How my personal journey illustrates what is possible when we
prioritize expanding access to the tools necessary to prepare
students and workers for AI
2. SeedAI's grassroots work provide AI skills to students
across the country
3. How SeedAI is working to ensure all Americans can benefit
from AI
1. My Journey
I always harbored a deep passion for technology and sports.
However, after graduating high school, I faced several obstacles that
led me on an unexpected path. Eventually, I found myself in Houston,
Texas, working in the oil and gas industry. When COVID-19 hit, I was
laid off, but this setback sparked a pivotal career change. In the fall
of 2020, I decided to pursue my interest in artificial intelligence by
joining the AI program at Houston Community College (HCC).
While at HCC, I immersed myself in the field by joining the
Computer Science Association to find community during the pandemic. My
dedication was recognized when I received the Pepsico Uplift
Scholarship and was named President of the Computer Science Association
club. Driven by a desire to create opportunities for others, I worked
with a group of hungry students and faculty to initiate the HCC AI &
The Future Conference, which has since evolved into the Gulf Coast
Conference on Artificial Intelligence. My efforts culminated in
becoming the first graduate of HCC's Associates in Applied AI Degree--a
first not just for the college but for the entire state of Texas.
This achievement, winning the Intel Global AI Festival National
competition, and presenting the project to Intel CEO Pat Gelsinger,
paved the way for my current role at SeedAI, which I joined in 2023.
2. Grassroots Work with SeedAI
My grassroots work at SeedAI has revealed key lessons about how we
should think about better preparing the American workforce for AI.
First, students from community colleges, Historically Black Colleges
and Universities (HBCUs), and other underrepresented demographics are
hungry and capable of harnessing the power of AI to their advantage.
Second, when presented with new perspectives and opportunities,
students from these demographics thrive in environments that require
resourcefulness and creativity.
Enabling greater participation in AI means more than just making
the technology perform more effectively for diverse populations. It
means ensuring that the trajectory of the technology bends toward the
needs of the many, not the few, and that all Americans can have a hand
in shaping the future of AI.
Hack the Future
Hack The Future (HTF), a flagship initiative of SeedAI has made
significant strides in engaging students and technology enthusiasts
across the Nation. To date, we have successfully hosted nine Hack The
Future events, each designed to inspire and educate participants about
the potential of AI and emerging technologies. These events have
collectively reached over 250 students, providing them with hands-on
experience and valuable insights into the world of technology and
innovation.
Our impact extends beyond traditional educational settings. At
DEFCON 31, the world's largest hacker convention, our HTF event engaged
over 2,200 users, showcasing the broad appeal and relevance of our
program. We've also organized specialized events such as Hack The
Future: Greenwood, in collaboration with Black Tech Street and the
White House Office of Science and Technology Policy, and Hack The
Future @ SXSW 2024, where we facilitated the participation of 47
students from Houston Community College, the University of Houston, and
Texas Southern University. Other notable events include Hack The Future
Cleveland, which partnered with Case Western University and JOLTI Law
Journal, and Hack The Future: Atlanta, collaborating with the Dream
Machine. Looking ahead, we're excited to announce our tenth event,
scheduled for October 11 in Houston, continuing our mission to
democratize access to AI education and opportunities.
Public AI Red-Teaming
Red-teaming has been a key tool in security research for decades
and typically requires specialized skillsets. \1\ But this does not
have to be the case with AI. With large language models, where the
interface is typically plain language, the barriers to participating in
red-teaming are dramatically lower. And when it comes to identifying
shortcomings in model performance, the larger and more diverse cohort
red-teaming a model, the better.
---------------------------------------------------------------------------
\1\ https://cset.georgetown.edu/article/what-does-ai-red-teaming-
actually-mean/.
SeedAI saw this as an opportunity to pilot the idea that modern AI
opens up opportunities for broad, diverse, and relatively unskilled
folks. In early 2023, SeedAI began working with cybersecurity and AI
experts at Humane Intelligence, the AI Village, the White House Office
of Science and Technology Policy, the National Science Foundation, and
the congressional AI Caucus to set our sights on an ambitious, large-
scale public red-teaming event at the AI Village at DEF CON 31. \2\
---------------------------------------------------------------------------
\2\ https://www.hackthefuture.com/news/ai-village-at-def-con-
announces-largest-ever-public-generative-ai-red-team.
As something like this had never been done before, SeedAI got to
work to design what a public AI red team would actually look like in
practice. SeedAI held the first pilot of a competitive AI red-team with
Houston Community College students at South by Southwest in March 2023.
\3\ SeedAI then hosted a second pilot at Howard University in June to
teach Howard students how to expose bias, potential harms, and security
vulnerabilities in generative AI models. \4\ With DEF CON scheduled for
just a few weeks away in August, SeedAI and our partners secured the
participation of Anthropic, Cohere, Google, Hugging Face, NVIDIA,
OpenAI, and Stability to provide model access to red-team participants.
---------------------------------------------------------------------------
\3\ https://www.youtube.com/watch?v=nida-w1J9r4.
\4\ https://www.hackthefuture.com/news/hack-the-future-at-howard-
university.
With all of the major AI players involved, the AI red-team at DEF
CON ended up being the largest ever generative AI red-team (public or
private) by an order of magnitude with 2,244 participants. \5\ SeedAI
and community partners Black Tech Street and Houston Community College
also brought 220 students and entrepreneurs from Tulsa, HBCUs, and
community colleges to DEF CON to join in. Our goal was to ensure that
the exercise would create a pathway for people from all walks of life
to participate in shaping AI while demonstrating that including more
diverse perspectives in red-teaming makes AI more resilient, reliable,
and trustworthy for all Americans.
---------------------------------------------------------------------------
\5\ https://www.forbes.com/sites/thomasbrewster/2023/05/04/biden-
white-house-backs-biggest-ai-hacking-event-with-google-and-chatgpt/.
The lessons learned from the DEF CON AI red-team highlight that
public red-teaming--while not replacing the need for robust internal
testing and evaluation practice--is a valuable addition to the broader
AI safety landscape and can shine greater light on harms related to
demographics and lived experiences. \6\
---------------------------------------------------------------------------
\6\ https://drive.google.com/file/d/1JqpbIP6DNomkb32umLoiEPombK2-
0Rc-/view.
While red-teaming is a natural entry point for greater
participation in AI, policymakers, industry, academia, and civil
society should seek to lower barriers to public participation in AI
through as many domains as possible.
Microgrants for Direct Student Empowerment
As part of our mission to democratize access to artificial
intelligence education and opportunities, SeedAI has implemented a
robust microgrant pilot program to fill gaps we saw in the AI skills
landscape. This initiative explicitly targets students and clubs at
community colleges, HBCUs, and other minority-serving institutions
across the United States. The program aims to provide crucial financial
support to those who might otherwise lack access to resources in the
rapidly evolving field of AI.
In a significant step toward realizing this goal, SeedAI forged
formal partnerships with nine community colleges and universities.
These partnerships create pathways for students to access a wealth of
opportunities and resources in the AI domain. To date, SeedAI has
distributed over $90,000 in microgrants across nine educational
institutions, directly investing in the future of AI education and
ensuring that a diverse range of students can participate in and
contribute to the AI revolution.
The goal of these microgrants is to pilot a new model for lifting
up students by lowering barriers to accessing AI skills. We are
incredibly proud of this work, but we recognize that this is not a long
term solution to addressing the needs of students across the country--
to ensure we can reach as many people as effectively as possible, we
are shifting our efforts to support national consortia. As we work with
other groups and initiatives to scale our impact, we will continue to
iterate on and prioritize the microgrants program and bring the lessons
we learn to inform and work on the national stage.
3. Ensuring AI Benefits All Americans
SeedAI works to ensure that everyone across the geographic and
economic spectrum--not just a select few--can influence and benefit
from transformative AI technologies. We believe that this will lead to
a fairer society, better AI policy, and a more robust and diverse
workforce, enabling the United States to maintain its leadership
position in driving technological innovation.
Expanding access to AI opportunities and preparing the American
workforce for AI will be critical to America's AI competitiveness
overall. Currently, the AI talent is concentrated in a handful of
companies primarily located in Silicon Valley, and many talented
researchers, workers, and entrepreneurs do not have access to the basic
infrastructure and resources they need to develop and test their ideas
and apply AI in their careers.
Infrastructure
Lowering the barriers to acquiring AI skills is just one piece of
this puzzle--skills alone will only go far. We need to invest in
creating the infrastructure so that all who wish to can access the
computational and data resources, testbeds, software, and tools
necessary to meaningfully engage and co-create with AI. This
infrastructure is what will enable us to move beyond just providing
workers with access to AI to actually democratizing participation in
shaping the technology.
To that end, SeedAI has strongly supported the creation and
complete funding of the National AI Research Resource (NAIRR). The
NAIRR pilot, launched in January 2024, is a proof-of-concept for the
eventual full-scale NAIRR, which will create national infrastructure to
connect U.S. researchers with the compute, data, and tools necessary to
conduct valuable AI research. The NAIRR pilot is led by the National
Science Foundation in partnership with 12 other Federal agencies and 26
non-governmental partners--an excellent example of the kind of public-
private partnerships we desperately need to create the robust public
infrastructure necessary for large-scale participation in AI.
Research from the NAIRR can be a powerful foundation of technology
that can be applied and tested by people around the country. However
the NAIRR itself is not yet fully funded. The bipartisan CREATE AI Act
would formally authorize the creation of the NAIRR, which is a critical
first step. \7\ To be fully effective, The NAIRR Task Force estimated
that the NAIRR will require $2.6 billion over a 6-year period. \8\
---------------------------------------------------------------------------
\7\ https://www.Congress.gov/bill/118th-congress/senate-bill/2714.
\8\ https://www.ai.gov/wp-content/uploads/2023/01/NAIRR-TF-Final-
Report-2023.pdf.
---------------------------------------------------------------------------
The National Applied Artificial Intelligence Consortium
In 2021, NSF Director Sethuraman Panchanathan laid out a vision for
the future of science in the U.S. that tapped into the ``missing
millions''--people who are capable of succeeding as scientists and
engineers but do not have access to pathways that lead into those
careers. \9\ We believe community colleges are the key to unlocking
these pathways.
---------------------------------------------------------------------------
\9\ https://www.aaas.org/news/nsf-director-lays-out-vision-future-
us-science.
This is why SeedAI is proud to partner with the National Applied
Artificial Intelligence Consortium (NAAIC), an initiative launched by
Miami Dade College (MDC) in partnership with Houston Community College
(HCC) and Maricopa County Community College District (MCCCD). The
Consortium's mission is closely aligned with SeedAI's goals of ensuring
broad participation in AI development and governance by building local
AI hubs and engaging diverse communities. The founding schools all
---------------------------------------------------------------------------
received microgrants from SeedAI in 2024.
The Consortium is a game-changer, not just for community colleges,
but for the entire AI-driven economy. Applied AI skills are the key to
unlocking the benefits of fundamental AI advancements, as well as well-
paying jobs and sustainable careers for all Americans. We can have all
the AI research in the world, but it doesn't make a difference if we
don't have communities prepared to apply the technologies to the
benefit of everyday Americans.
Tulsa Hub for Equitable & Trustworthy Autonomy
To that end, SeedAI is proud to support the Tulsa Hub for Equitable
& Trustworthy Autonomy (THETA), which was recently awarded $51m by the
U.S. Economic Development Administration (EDA). \10\ SeedAI will work
with Black Tech Street to provide strategic guidance and capacity
building for the AI/AS Program for Innovation, Research, and Education
(ASPIRE) initiative. This includes procuring high-performance computing
resources and organizing underlying projects for community benefit.
---------------------------------------------------------------------------
\10\ https://www.eda.gov/funding/programs/regional-technology-and-
innovation-hubs/2023/Tulsa-Hub-for-Equitable-Trustworthy-Autonomy.
We view partnerships like the Tulsa Hub and the NAAIC as critical
``last-mile'' components of the NAIRR. These partnerships address the
critical need to ensure that research from the NAIRR can be co-
developed and leveraged by communities across the country.
Legislation to Increase Participation in AI
SeedAI also supports bipartisan legislation like the Expanding AI
Voices Act, which aims to diversify AI research and workforce
development. \11\ This bill would support capacity-building in Minority
Serving Institutions (MSIs), HBCUs, and Tribal Colleges. By expanding
the National Science Foundation (NSF)'s ExpandAI program, the act
ensures that the future of AI is inclusive and representative of all
communities.
---------------------------------------------------------------------------
\11\ https://www.Congress.gov/bill/118th-congress/house-bill/9403/
text/ih.
Another is the bipartisan Workforce for AI Trust Act, which aims to
develop a multidisciplinary workforce that can advance the creation and
deployment of safe, fair, and transparent artificial intelligence
systems. \12\ These proposals can leverage the lessons from work like
ours for the broad benefit of Americans.
---------------------------------------------------------------------------
\12\ https://www.Congress.gov/bill/118th-congress/house-bill/9215/
text.
---------------------------------------------------------------------------
Conclusion
There is both incredible appetite and untapped potential in
American workers and students across the country to participate in and
benefit from AI. SeedAI is eager to work with Members of this
Subcommittee to help realize the vision of preparing the American
workforce for AI and leverage the lessons we've learned from working
with students across the country. There are several key opportunities
we think would help make this happen as fast and as equitably as
possible.
First, we are eager to continue expanding our Hack the Future
program to serve communities in your states and across the country.
Please don't hesitate to reach out if that's of interest.
Second, we strongly support passage of the CREATE AI Act to
authorize the NAIRR, and support its full funding of $2.6 billion a
year for 6 years. We also support passage of the bipartisan Expanding
AI Voices Act and the Workforce for AI Trust Act. The development and
application of AI is an ecosystem, and we need to support each critical
component.
Third, as one of the first graduates with an Associates Degree in
Applied AI, I encourage Members of the Subcommittee to explore
opportunities to eliminate counterproductive education requirements for
entry level jobs related to AI. Skills are what matter, and creating
programs to cultivate these skills in the American workforce only go so
far when many jobs require 4-year degrees that prevent qualified
workers from accessing these opportunities.
Fourth, most importantly, we encourage the Subcommittee to leverage
its power of attention. Progress on this topic requires robust public
private partnerships and tapping all stakeholders in the government,
industry, and academia to make changes to facilitate greater
participation in and access to AI opportunities.
By promoting these opportunities where they exist, incentivizing
their creation, and providing them with the resources and
infrastructure they need to be successful, we can transform the AI
ecosystem to be one that is more inclusive and participatory, more
geographically diverse, and more beneficial for all Americans.
______
Senator Hickenlooper. Terrific. Thank you, Mr. Wilson. Let
me start the questioning--am I missing anything? I always look
to the Ranking Member because he really actually understands
politics at much--at a deeper level than I do. When I am in
trouble, he kicks me under the desk, or nudges, I should say.
Anyway, Mr. Kotran, aiEDU is launching AI literacy
initiatives across the country including a recently launched
partnership in Colorado. How do you work with local leaders and
employers to help customize the training resources and ensure
that they are responsive to local or regional training needs?
Mr. Kotran. Yes, thanks for the question. So our work is
really different state to state. Colorado is an example of a
state where we have actually found the opportunity to push
forward a statewide initiative.
In Colorado, we have worked with an organization called the
Colorado Education Initiative, which actually aiEDU worked
directly with school districts and spun out into an independent
nonprofit. And we have essentially plugged in as a subject
matter expert for CEI. CEI over the past year has convened
hundreds of stakeholders across the education system, the
workforce, and the political community.
This is really top down advocacy work to really create the
demand and the awareness for not just the goalposts, which are
how do we make sure that teachers are trained, and students are
ready, but also creating awareness about what are the systemic
components that have to be in place, which ultimately comes
down to training and building capacity with teachers.
Senator Hickenlooper. Got it. Thank you. Mr. Wilson, as we
undergo this new technological revolution, we want to make sure
that workers aren't left behind.
How do you think the Federal Government can best position
training resources to make sure that they reach traditionally
underserved workers, including those in rural communities,
those who might speak different languages, those completely
outside the technology industry?
Mr. Wilson. Thank you for that question, Chairman. I look
at it from the perspective of--I am looking at it from
community college perspective. I think community colleges
across the country are the most representative of communities
and constituencies.
I think if we put focus on those 2 year programs and
actually doing workforce related programs and putting funding
toward those programs, it more so prepares students to be ready
to work right after they get that 2 year degree instead of the
traditional, you transfer from a 2-year and go to a 4-year and
then try to get a career.
A lot of students at the community college level are
actually equipped. And if there is just more funding put toward
that, I think it would be very beneficial for students coming
with a 2-year degree to get into the workforce.
Senator Hickenlooper. Great. Thank you. Mr. Meyer, as some
employers have been experimenting with AI applications, we have
seen that creating opportunities for workers to provide
feedback is important.
What steps are you taking, or should others be taking, to
collect feedback from workers about their experiences as they
interact with these AI applications? And how would that
feedback impact your efforts going forward?
Mr. Meyer. Thank you very much for the question, Senator
Hickenlooper. First of all, employers, as they are doing, as
most of them are doing right now, need to have their open door
policies for their employees to come in and their workers to
come in to tell them like, this is what my experience has been
in this because we are going to--because employers then should
take that feedback and use it to improve the employee
experience.
Because just from a personal standpoint, as far as I see,
the biggest value for AI is for people and organizations to
focus on what is really important. Yes, WRI, what is really
important.
Whatever your what is really important is or whatever
organization you are at, using AI and having people who can use
it will give them, give those individuals the time, the
organizations, the time to focus on their WRI, what is really
important.
You are never going to know what is really important unless
you are speaking to your people.
Senator Hickenlooper. Great. Thank you. Dr. Kimbrough, we
know that AI has the potential to impact workers across all
industries, all levels of experience. What are some of the most
in-demand technical and non-technical AI skills being sought
out by employers?
Does LinkedIn's data mirror training participation for
those skill sets or those workers across industries? Are they
seeking out such training opportunities and are there perhaps
some industries that are more engaged in training than others?
I will say it nicely.
Dr. Kimbrough. Thank you, Chair Hickenlooper. I would say
that in terms of the most in-demand skills that we see both for
like technical talent and non-technical talent, starting with
technical, it is very clearly things like generative AI,
machine learning, generative neural networks, things like that
are highly technical, and these are used by engineers who are
forming and building these models and training them.
The non-technical AI skills for the average person who
wants to build their literacy is more around use of tools in
the workplace or at home. It could be things like Copilot or
ChatGPT. So just basic applications. But if you go broadly and
say, what is LinkedIn doing? I think that was another part of
your question, is how do we think about what we see in our
data.
There are industries that are adopting it more quickly than
others. We do see industries such as retail, administrative and
social--sorry, administrative services. And also I would say
actually things--think about things like even in hotels,
accommodation, restaurants, they are all starting to adopt
generative AI tools.
Senator Hickenlooper. Interesting. All right, I am going to
turn it over at my 5 minutes, but I will be back. We get to
have second rounds, I suspect.
Senator Braun.
Senator Braun. Thank you, Mr. Chairman. Mr. Meyer, when I
look at technology as it came on the scene decades ago, it
scared certain generations. The ones that grew up with it
seemed to embrace it, and it was like a second language or
second nature.
When you compare that to AI currently, talk a little bit
about the generational difference. In other words, kids that
have grown up with technology, do they view AI as just a new
horizon? And for the folks who didn't grow up with it, how do
you explain this to them?
Mr. Meyer. You are right, it is--some of this stuff can be,
in fact, really scary. I am just thinking about when my mom had
to use a computer for the first time when she was working in
health care.
Because that is how I got my job in health care, just like
a lot of people do. Their mother or father worked in health
care and that is where you wound up. But I have to tell you, I
think people are excited about this technology.
Something to think about, the most amazing thing about AI
is that we look at it today and this is the worst it is ever
going to be. It is only going to get better from here, and it
is only going to get more advanced from here.
When we are looking at it now, in like three or 5 years
down the road, it could be completely different, but you are
right. We have been dealing with multigenerational workforces
in HR for the last two decades, and this is certainly something
that we are going to have to focus on to alleviate people's--
alleviate people's anxiety when it comes to AI for those who
may not be as tech savvy.
But also on the flip side of that, it is also a great
reverse mentorship opportunity for organizations where you have
younger generations in the workforce actually mentoring older
generations of the workforce. It is actually pretty cool to
think about stuff like that going on.
That is something where we at SHRM and my colleagues who
are in the HR profession, we have been dealing with this for
the last--like I said, definitely for the last two decades. And
a lot of us are looking forward to the opportunity.
Senator Braun. Thank you. Mr. Wilson, would you want to
weigh in on that as well, and with emphasis especially on the
generation that is still trying to learn technology--has been
thrown in their lap. They didn't grow up with it.
Mr. Wilson. Yes, a great question, Senator. From my
perspective, I grew up in the generation with technology,
obviously. So this isn't really a new phenomenon. We grew up
watching many robotic movies and AI taking over the world, so
it is kind of something we have been primed for.
Whereas my parents and my uncle's generation, they kind of
were more uneasy about it. And I think that is where the AI
anxiety portion that I talked about previously comes in because
we know there is a lot of anxiety around what the possibilities
AI are.
I think it is important that both parties, all age groups,
all generations get a clear understanding and foundational
knowledge of how AI can be beneficial, but also the risks. And
I think just with the older generation, we have to push that
message a little more just because it is something they are not
used to.
But I do think that message needs to be consistent so we
can get to the injury portion where people aren't so afraid of
what AI can do.
Senator Braun. Thank you. Dr. Kimbrough, recently, we have
seen some states, localities attempt to implement policy in the
AI space. New York City enacted a law regulating employers use
of automated employment decision tools in hiring and
promotions, requiring employers to audit their HR Technology
systems for bias and publish their results.
In California, the Governor signed an Executive Order
directing state agencies to adopt a proactive approach to AI
regulation, and to leverage the state's procurement power to
promote so-called trustworthy AI principles.
Are we at the point where we know what trustworthy AI
principles are? And based upon what you have seen early out of
the gate, what do you think of actions like we have seen here
in New York and California?
Dr. Kimbrough. Thank you, Ranking Member Braun. Well, I
would say from my perspective as an economist at LinkedIn, I
think we are all searching to commit ourselves to an area where
we have the most responsible AI principles. At LinkedIn, this
is exactly how we operate.
We seek, for example--what I can speak to, we seek to put
our members first. So, where we can uphold trust, where we can
promote fairness and inclusion, where we can be extremely
transparent, those are the kinds of principles that we adhere
to in our own work at LinkedIn, to put members first and kind
of create this vision of inclusion.
I think it also empowers people when they understand how it
is being used. They feel more empowered and there is more trust
there. Thank you.
Senator Braun. Thank you. And Mr. Kotran, so much of this
arena is a competition with other places that will find a way
to use it more productively, or on the other side of the
spectrum, maybe come at you with it from a less wholesome point
of view. What do you worry about?
How do you see kind of the way we handle it vis-a-vis what
other countries are doing, and you can see where they would
want to probably be on the leading edge of it ideally to make
their economies more productive.
Or do you fear that this will be something we are going to
battle similar to what we do with the malfeasance that we have
got in our own technological world where you have got folks out
using it with ransomware and other ways where they would use it
as a negative tool. Where do you see AI on that kind of a pivot
point of good versus evil?
Mr. Kotran. Yes, I think I mean, you mentioned----
Senator Braun. I saved the best question, I think, for you,
Okay.
[Laughter.]
Mr. Kotran. Thank you, Senator. Look, there is a lot of
examples of bad ways that bad actors can use AI. I think, when
I look at what other countries are doing relative to the U.S.,
I am concerned about the fact that we are really on the back
foot in terms of being ready for the economic displacement that
is going to come from these tools.
Like make no mistake, AI may not be coming for teachers
jobs. It is coming for doctors, lawyers, designers,
mathematicians, engineers, data scientists, paralegals,
assistants, customer service managers, marketing managers,
advertising managers. These are all careers that the University
of Pennsylvania identified as being nearly fully exposed to
large language models over the next 5 years.
Again, this is not necessarily going to happen gradually
and comfortably. It is going to happen all at once. If you are
a company with $5 billion--$5 billion in revenues, your
revenues are down--your profits are down $1 billion in the
middle of a recession and your margin is 20 percent.
You can increase sales by $5 billion, or you can cut costs
by a billion. Companies are going to cut costs. They are going
to use these tools to make up the difference as they lay off
workers. And what I am concerned about is the fact that in the
U.S. we do not have a plan to make sure that we are talking not
just to the current workers, because that is a huge issue, but
also to students who are making big decisions about what career
path do they go.
Which college career do I invest in? Should I spend
$200,000 to go and get a law degree or a computer science
degree? And you have the smartest experts in the field saying
those jobs are going away. And at the same time, that
conversation is happening in places like this in Silicon
Valley. It is not happening in schools, and that is really
concerning to me.
Senator Braun. Thank you.
Senator Hickenlooper. I think one of those vulnerable jobs
might be that of a U.S. Senator. You never know.
Senator Kaine.
Senator Kaine. Thank you, Mr. Chairman. And thanks to the
witnesses. So, if I am a 30 year old Virginian watching this, I
would--wow, I should become AI literate. I should improve my
employability and future career by gathering skills. But it is
not just about getting skills. It is about getting skills that
come with a credential.
Whereas if I moved to Virginia, to Vermont or to Vegas, I
am bringing a skill, but I am also bringing a credential that
an employer will recognize, Okay, yes, you have got the skill.
Are there industry standard recognized credentials in the AI
field that are portable like that where I could gain their
credential and move somewhere else in the United States and an
employer human resources professional to look at it and say,
yes, that--Okay, we understand that. I see you shaking your
head no, Mr. Kotran.
Mr. Kotran. I think the--thank you. Senator. There is often
this reflexive reaction to say, well, yes, can we create an AI
credential? What are the AI jobs of the future? And when you
talk to folks, some people say, well prompt engineer is the job
of the future.
To me that is as ludicrous as going back to 1998 and saying
a Google search engineer is going to be a job of the future. We
do not know what the jobs of the future are. If you are 30, the
real question is, how are you going to build the soft skills,
the critical thinking, the ability to work in teams
collaborative----
Senator Kaine. You call those durable skills in your
opening testimony. I like that better than soft skills.
Mr. Kotran. I think that is right. And I think that is
actually a really big area that we need to focus on, is how do
we actually codify what these skills would look like,
especially in a world where the workforce is going to change
faster than these credentials and frankly our education system
is going to be able to change.
This is a really big blind spot that we have. This is like
a huge tanker ship we are trying to now navigate down River
Rapids. And the best answer that I can come up with is the
teachers, the professors, the schools, the folks who are really
thinking about training our teams, they are on the front lines.
To me, it is really scary the fact that a company can
depreciate an investment in an AI tool and get a tax--a
favorable tax treatment for that. They can't depreciate an
investment in employee upskilling.
Senator Kaine. Yes, yes. Dr. Kimbrough, you used the
phrase, we should expand skills based training, in your
testimony. So obviously we are moving from a world where it is
about degrees, to degrees are great, but also skills
credentials, as long as they are valid, are great. What did you
mean when you said things we can do is to expand skills based
training.
Dr. Kimbrough. Thank you, Senator Kaine. When I was talking
about skills based hiring in particular, if I may use that
term, I was thinking a lot about the data that we see at
LinkedIn, which if we think about what are required skills for
a job and push away the idea of like what credential do you
have, even if it was a degree, a former title, prior
experience, and you say, what skills are you able to perform?
What we find is that for any role on average, you can
increase the pool of eligible candidates sometimes by as much
as 20 times. And so, if I may, I would just say what I think
can happen is ideally we would be promulgating more pathways
for employers, for recruiters, for state, and local, and
Federal Government roles to be skills based as opposed to,
let's say, degree based or prior experience, because you pull
in many more people.
I would just add one last point, which is that what we
found in our own data is when you expand that talent pool by
thinking about what people either can do or potentially can do
as a skill, you actually include more underrepresented people.
For example, women are 20 percent--26 percent more
represented when you look at a skills based pool than a non-
skilled based pool.
Senator Kaine. We have a lot of work to do here in Federal
policy. Senator Braun and I are working on this. We think Pell
Grant should not be just limited to college degrees. Why not
high quality career and technical education? And yet we have
had a long bias against career and technical education.
It should be high quality, but why would we benefit college
and not that? The military has tuition assistance programs for
active duty. My son used to be a marine officer. He could
approve somebody in his platoon for up to $4,500 bucks a year
for a college class.
But if he had an ordinance specialist who was trained like
in welding and all he needed was $300 bucks to take the
American Welding Society certification exam, which is a
credential that is completely portable, he would have to say,
no, I can't give you $300 because it is not college.
Mr. Wilson, I will finish with you if I can real quick.
Community college. I think this is a key part of this. And I
mentioned a lot of the folks who gain these skills at a
community college. They are there to gain the skill, not
necessarily the associate degree.
Sometimes the best way to gain the skill is in a short
term, high intensity class that isn't a 15 weeklong semester
and hence is not eligible for Pell Grant. Am I right about
that?
Mr. Wilson. I believe our program was available for Pell
Grant, Senator. Now, what I will say is the workforce element
at the community college level, I think is what really--we
should really pay attention to.
It is a 2-year format where you are actually working on
what you are going to be doing once you graduate instead of
taking that first 2 year to do general education classes. You
are actually doing the hands on work, and that is why the
applied in the Applied degree comes into play because you have
actually applied the skills to something.
Whether it is in your last class that you have to take or
whether it is an internship you got to do over the summer, you
have actually experienced that and you are more equipped to
actually get into the workforce. And I think if we pay
attention to these degree programs, there are a lot of
community colleges around the country.
Houston Community College is a great example, Miami Dade
College, Chandler Gilbert Community College. These are
community college that took the initiative to start AI
workforce programs at their school, and now some of them have
been able to turn it into a bachelor's at the community college
level, which many people didn't know was possible. So I think
it is very beneficial.
I have seen the students coming out of these programs. I am
one of them. A lot of them are very well equipped. A lot of
them are very sound in what they know, and they are very
hungry. They just need--many companies maybe change their
hiring practice or human resource departments to take off the
master's or bachelor's degree requirement because they know
they are already capable with their 2 year degree that they
got.
Senator Kaine. As I hand it back to the Chairman, we are
looking at the same thing with Federal hiring. Do you have to
have a college degree to get a lot of Federal jobs? And the
answer is no. Do you have to have a college degree to be
assigned by a Federal contractor to work on a contract with the
Federal Government? If you have the skills to do the job, it
shouldn't be restricted to those with college degrees when we
are in a skills based world these days. Thank you. I will yield
back.
Senator Hickenlooper. Yes, I couldn't agree more. We did,
back when I was Governor of Colorado in 2017--I guess 2016 and
2017, we went back and looked at pretty much every job that we
offer and basically downgraded almost half from saying you had
to have a college degree, to saying what you really need to be
successful in this job.
Oftentimes we would put the college degree in there because
it was part of that emblem of achievement that made people feel
that the person was responsible and a good worker in some way,
whereas that really was no reflection of the skills necessary
for that job. And I think that is a good point, obviously.
Senator Budd.
Senator Budd. Thank you, Chairman. I thank the Ranking
Member as well. And I appreciate what each of you all bring to
this conversation as we wrestle with this in the Senate and
Congress in general.
I think this is an issue which I think in your term, Mr.
Wilson, brings anxiety for many, and eagerness and anticipation
for a lot of others. So I am on the optimistic side, but though
I acknowledge the whole spectrum here.
I am a believer in the promise of innovation here in the
U.S.. I think that the advancements in AI are going to change
how jobs are done, some of you mentioned that, and not
necessarily eliminate them in total.
Dr. Kimbrough, given LinkedIn's unique insights into
employment trends, what kind of changes do you think AI is
going to make to the types of jobs, not that are going to be
eliminated, but the ones that are going to be in demand over
the next 5 to 10 years?
Are there specific industries or sectors where you foresee
job growth due to the advancements in AI?
Dr. Kimbrough. Thank you, Senator Budd, for that question.
It is a really hard question, actually, and I am sympathetic to
it, Mr. Kotran said----
Senator Budd. We are betting on what you have to say.
[Laughter.]
Dr. Kimbrough [continuing]. Nobody--I recognize that. I am
starting to sweat here. Yes, what I would say--well, one I
would say is, it is really hard to know what the jobs of the
future are. I will give you an example of why I say that. Every
year, LinkedIn puts together a list of the fastest growing jobs
that we see on our platform, and there are millions of jobs on
that platform.
This year, two-thirds of the jobs on that list did not
exist 20 years ago. So I am very humble as an economist that it
is hard to forecast. But I want to answer your question, so
what I would say is some of the things that I think are likely
to be actually most in demand are some of what we are calling
the durable skills or the human skills.
Oftentimes people ask me, what should my child be learning
or studying? And I am like, abstract problem solving critical
thinking, communication. Some of these skills are the most in-
demand skills every year on our platform, communication,
collaboration, leadership strategy. So these are the things
that I don't expect AI to replace. Thank you.
Senator Budd. Sure, sure. Maybe some of the community
college skills. I was thinking of the heat generated through
some of this computing, and that involves a lot of plumbing and
processing and HVAC. So there is a lot of skills like that
perhaps.
Dr. Kimbrough. We do actually--sorry to interject. We do
see a lot of demand for electrical engineers, electricians,
mechanical engineers, plumbers, truck drivers.
Senator Budd. Yes, absolutely. In a landscape where AI is
becoming more widely used in the workplace, it is more
important than ever to be able to quickly reskill and upskill
workers, as many of you are talked about.
That is why I introduced both the Pell Act and the Employer
Directed Skills Act. The Pell Act would allow Pell Grants to be
used for short term programs. I think like the welding program
that you mentioned.
This would allow job seekers to enter the workforce faster
and make their skills align with the current needs in an AI
driven world. The Employer Directed Skills Act will give
employers more resources to prepare job seekers for open jobs.
This is another way to make sure that we are leveraging
workers to fill changing roles as AI changes the demands of the
workforce. Dr. Kimbrough, can you share how you think that some
of these legislation, not just these but maybe some others
even, will help job seekers and employers prepare for jobs that
are going to change as a result of AI?
Dr. Kimbrough. Yes. So from our data at LinkedIn, what we
see is there is a huge scarcity of AI literacy talent or AI
technical talent. There just isn't enough. There is much more
demand than there is supply for this. And I don't think that we
are going to replicate or close that gap very easily.
It is just really critical to have as many pathways as
possible to create opportunities for America's workforce to
reskill and upskill in a continuous fashion, not just a one
off. I go to a community college, and I am done, or but
really--and not to put down the idea of community college. Very
supportive of it.
But really opportunities of small learning. And so one of
the things that I would say is we very much support anything
that extends the pathways, even if there is the short term
programs where people can quickly get credentials to be ready
to work in the workforce, to meet this--the scarcity, this huge
demand for AI literacy that isn't being met.
Senator Budd. Thank you very much. One comment--several
comments I wanted to make. I know your parent company,
Microsoft, recently announced a deal to restart--this is really
outside of maybe some of your projections, but they restarted
the Three Mile Island nuclear power plant in Pennsylvania.
The power plant would develop data centers which have a
growing demand as AI develops and is used more widely. Just one
query in ChatGPT is estimated to use 10 times as much energy as
a simple Google search.
For the first time in decades, our power resource--our
power requirements are going to increase. But I am very
concerned. I mean, I see articles like in today's Wall Street
Journal and it continues on about the need for more clean
energy, or the steps to get to clean energy. So I am very
concerned about the current Administration and their positions
in regards to clean energy. They don't support it.
Vice President Harris, she has been on the record
supporting banning fracking. A lot of that out of Pennsylvania,
which has allowed us to be a net exporter of energy. We need
all forms of energy, including natural gas, which has 50
percent fewer carbon emissions than coal. Our biggest
competitor, China, is going to be using.
We are not talking that they are going to be using clean
coal technologies either. And you can bet that China will make
sure that they have the energy that they need to surpass us in
AI. So as we think to the future, in the near future, I hope
that we keep that in mind regarding our energy needs here in
America. Thank you.
Senator Hickenlooper. Thank you, Senator Budd.
Senator Markey.
Senator Markey. Thank you, Mr. Chairman. And thank you so
much for having this very important hearing. We have heard
extensive discussion today about the importance of educating
and training the workforce for the increasing use of AI, but we
must also prepare the workforce by protecting workers from the
potential harmful and discriminatory impact of workplace AI
tools.
Over the past decade, companies have increasingly
integrated AI into their hiring. Their firing, and their worker
management. And time and time again we have seen how the biased
and discriminatory algorithms powered by AI are used to hire,
to assess, and to set working conditions that can hurt workers.
Here are some of the examples. Algorithms that match
candidates with job opportunities that refer more to men than
women for open roles. Generative AI chat bots that
systematically weed out resumes from Black men, from the LGBTQ
community. And AI that surveils call center workers and grades
their performance that are more likely to sanction Black women
for the tone of their voice.
AI, the sinister side of AI, that is unacceptable. To each
of the witnesses from left to right, yes or no, do you agree
that algorithms used to hire and make employment decisions
should be free from bias and discrimination? Dr. Kimbrel, all
the way down.
Dr. Kimbrough. Yes, at LinkedIn, we are committed to
upholding the principles of responsible AI--yes.
Mr. Kotran. Yes.
Mr. Meyer. Yes.
Mr. Wilson. Yes.
Senator Markey. That is why yesterday I introduced my AI
Civil Rights Act of 2024, comprehensive legislation that would
prevent discriminatory algorithms from being used in critical
decisions, including employment, because we cannot prepare our
workforce without protecting our workers simultaneously.
This technology is Dickensian. It is the best of
technologies and the worst simultaneously. It can enable. It
can ennoble. It can degrade. It can debase. It can do it all.
And we want the best, but we also build in safeguards to
protect workers from being discriminated against.
I would also like to take this opportunity to seek
unanimous consent to enter into the record a letter from the
National Nurses United, if I may, Mr. Chairman.
Senator Hickenlooper. Without objection.
[The following information can be found on page 49 in
Additional Material:]
Senator Markey. As the letter states, the development of
clinical algorithms, ``involves significant use of judgment by
their creators and creates the opportunity for creative bias,
from conflicts of interest, limited perspectives on the lives
of racial minorities, or implicit racial bias to be introduced
into the algorithm.''
That is why my AI Civil Rights Act would require companies
to test these algorithms for discriminatory impact before their
use, preventing biased AI from being incorporated into resumes
screening, job recommendations, worker performance, and even
health care decisions.
Make no mistake, we can have an AI revolution, while also
protecting the civil rights and liberties of everyday
Americans. We can support innovation without supercharging bias
and discrimination in our society. And we can prepare workers
while safeguarding their rights. And I might also add, well,
talking about the fact that a ChatGPT search consumes ten times
more electricity than a Google search.
Yes, we are going to need more electricity in our society,
but the industry is going to have to build in more efficiency
into their AI technologies. They are promising to find the cure
for cancer. Well, let's figure out how to make the industry
more efficient. Can't be having a search consume 10 times as
much energy as a Google search.
Do that work. Use AI, I guess, because you are promising to
find the cure for cancer. How about finding a way to do all
this much more efficiently? What are the algorithms you are
using to do that?
In addition, it is very disheartening for Donald Trump to
promise to repeal the IRA. To repeal the wind and solar and
battery storage technology subsidies that are transforming our
Country, which can make the use of electricity much more benign
in terms of greenhouse gases.
He did say to the natural gas industry, if they give him $1
billion, he would repeal the IRA. Well, that is the way in
which we can make AI compatible with the climate goals that we
have simultaneously.
I just, I heard an earlier comment, and I'll just throw
that in, that we just can't say it is going to be coal and
natural gas because we know that we have a renewable revolution
that is also happening, and we can't stop that. We can't slow
it down. We have to make sure that we move forward. Thank you,
Mr. Chairman, very much.
Senator Hickenlooper. Thank you.
Senator Casey.
Senator Casey. I have nothing for today. We can go ahead.
Senator Hickenlooper. Senator Braun--he has to run. Do you
have additional questions?
Senator Braun. Yes, I want to--since we didn't pass into
another question, I want to go back to the original or the
ending conversation we had.
Geopolitical competition to me is so important and we are
at that kind of point in time where there is so many things at
stake and we have got this new, and I think the point has been
made, we are going to see what opportunities and challenges are
created by it.
For anyone that would have a view on it, I would like to
know what areas of commerce, industry, and then what countries
have gotten out fastest using this. And has there been any
malicious use of it to date or has it all been beneficial?
I mean, because we watch this, we hear about it. Do you
keep track of it? Who has been out on the leading edge either
in an economy, a country, or a sector of commerce or industry?
We start over here, Dr. Kimbrough.
Dr. Kimbrough. Yes. So I really can only speak to the kind
of data that we see on our platform at LinkedIn, so I am really
speaking about talent trends here.
But in terms of building up capacity for skilling, I would
give one example, which I think is a very interesting one,
perhaps not scalable for the U.S., where in Singapore, for
example, there has been--the Government has actually provided
training funding for every single citizen of Singapore so they
can upskill in AI.
I know that is a smaller country than the U.S., but that is
one example. Another example I would just point to is we see
very fast growth in AI literacy and technical skills in
countries like India and Israel. And so, there are a range of
places where the technical training is just growing much
quicker than it is in the U.S.
Senator Braun. They are already ahead of us.
Dr. Kimbrough. Well, I would say the U.S. has the most--in
terms of level, the U.S. is very competitive. We have a third
of the AI technical workers in the U.S., that are here, and a
third of the AI literacy skills are workers in the U.S.. So we
have quite a bit, but the growth is faster in other countries.
Senator Braun. Mr. Kotran.
Mr. Kotran. I mean, really, the--I mean, this is a game
between the U.S. and China. I think we are still--as far as I
understand, we are still in the lead, although what we are
seeing from China is interesting. One of the things that they
just have an advantage over us is with respect to the actual
regulation of the technology.
One of the things that they recently announced is a
requirement for AI generated images to be tagged as such. And
this is happening at a time when in the U.S., there has been
rampant issues of students making use of notification apps,
which are basically AI apps that can take an image of somebody
and de-clothe them.
This is happening in high schools all over the country. And
the companies are based in places like Belarus, so we don't
have the ability to shut these companies down. And in the U.S.
and in the West in general we are hampered by the democratic
system and the fact that we have to sort of iteratively build
the regulations around as the technology evolves. There is
advantages and disadvantages to that.
Senator Braun. Mr. Meyer.
Mr. Meyer. I don't have any intimate knowledge of your
question. However, I do say at SHRM, one of the things we
believe in is AI, artificial intelligence, plus HI, human
intelligence, equals ROI, return on investment.
I think we are ahead, but I think other folks are catching
up. And I think it is really important. But one of the things
is that, and I just want to add this, is that when it comes to
catching up or staying ahead, one of the great things about AI
is the individuals have the ability through AI to exercise a
lot more of their own ownership over their own professional
development.
I think that is the really amazing thing about HI, is that
you can really own your own development. And that is how
organizations and individuals will be growing with this
technology, so.
Senator Braun. Mr. Wilson.
Mr. Wilson. Yes, thank you for that question, Ranking
Member Braun. I don't have any information in front of me today
to really give you a solid answer, but I would like to follow
with your office with some more information later.
Senator Braun. Thank you.
Senator Hickenlooper. If he leaves, then I have got you all
to myself. This is every Senator's dream. I thought it would be
interesting--and I won't keep you too long, I promise.
But I think that this notion of where do we get our
standards, how do we codify, what is the appropriate way to do
this, it ties back into that--a little bit of that disadvantage
that we have compared to totalitarian countries where they
could just decide that everything is going to be somehow
watermarked or that transparency on what is and what is not AI.
Personally, I think that is one of the--should be one of
our absolute imperatives with a greater sense of urgency. But I
would love to hear each of you just talk about where do you
think, how should we go about finding the standards, the
guidelines for this?
Dr. Kimbrough, I was also going to ask you just because I
am so out of touch. How big is--like, how many members do you--
you were talking about your members. How many members do you
have?
Dr. Kimbrough. Worldwide, LinkedIn has over a billion
members.
Senator Hickenlooper. Oh my goodness.
Dr. Kimbrough. They are not all online at the same time.
But yes, we have a billion members. In the U.S., it is just
about 200 million.
Senator Hickenlooper. Gosh, talk about energy usage.
Anyway, so what's your slant on this notion of how do we codify
standards? Where do we get standards and how do we codify them?
Dr. Kimbrough. Chairman Hickenlooper, I am a macro
economist, so I am not going to--I don't want to wiggle out of
the answer, but what I was----
Senator Hickenlooper. I am an out of work geologist and
they expect me to have an opinion on these things.
[Laughter.]
Dr. Kimbrough. I am sure you are way smarter than I am.
What I would say is I think that where we stand at LinkedIn is
really thinking about how do we embrace accountability for what
we do?
How do we ensure that we are transparent about what we do?
How do we make sure we put our members first, which is
paramount for us. And starting with those kinds of principles,
everything we build, every design choice we make is around
ensuring that we are providing value to the member first and
protecting them.
Senator Hickenlooper. Okay.
Mr. Kotran.
Mr. Kotran. Yes, I am also going to differ on talking about
specific policies. What I will say is, I think it is folly for
us to ever imagine that we can regulate away the risks that AI
poses.
I think there is like soft examples of like a worker making
a decision to use an AI tool to screen an applicant or even to
like write interview questions, and it would be very hard for
us to really think about what would even look like to have
regulation that completely blocks that.
Then there is also logistically things like a watermark, a
digital watermark. I mean, it just may be impossible for us to
truly block all the companies that are going to provide people
with the ability to create AI images, because many of them will
be overseas and out of jurisdiction.
But I will say that whatever the standards are, you have to
have an informed and educated populace and we have to actually
build social norms. And the same, the social norm, like nobody
has their phones out on the table.
That is something that when you go to dinner, people don't
sit and that you turn your phone over. I think there needs to
be a norm where someone shows you a generated image of somebody
and you almost respond like aghast, like oh my god, did you ask
permission to create their image.
I think right now, we haven't really quite figured out what
is the role of even just sort of these social norms or these
soft standards, let alone the hard ones.
Senator Hickenlooper. Yes. Although those norms are
dependent upon somehow some standards, some regulations in
place. And you--we know we are not going to get to everybody
who runs the stop sign, right, or slides through a yellow
light, runs a red light, right. But we still have those
regulations.
If we had watermarks, certainly we are not going to get
everybody, but once you set the rules, you are going to get 98
percent, and you will at least have an avenue to go after the
people that abuse the system.
Mr. Kotran. If I may, so I actually started my career in
AI, working on ethics and governance. And we worked with NIST
on how do you set standards for the use of AI in the legal
system.
Senator Hickenlooper. Well just parrot back some of that
stuff. Come on.
Mr. Kotran. Well, so what I would say is any conversation
about AI ethics, and you are going to hear the same things. You
are going to hear; algorithmic bias is something we need to
deal with. You are going to hear about how do we make sure
there is a human in the loop.
I have heard this rhetoric for now, six, seven, 8 years.
And so none of this is new. I mean, we actually know, roughly
speaking, what we need to aim toward. And others like very
tactical, practical policy questions of do we create laws and
enforcement mechanisms?
If not, then I think we have to answer the question of
like, well what are the other incentives that we can create,
whether it is for businesses or otherwise? But I just don't
know that we can--I don't think business standards by
themselves are sufficient.
Senator Hickenlooper. Right. Well, we will have--we can
discuss that more fully later. Mr. Meyer, what is your stance
on some of these--this notion of standards?
Mr. Meyer. I guess the standards that are going to be set
essentially are going to come from you, Senator Hickenlooper,
and others like you. But here is the thing, and this is the
most important thing, there needs to be a balance between
responsible use. And responsible use should be paramount.
These tools need to be developed, they need to be used
without bias. And they need to be developed, and they need to
be used without hallucinations. Those are the two big things.
However, we also need to not have rules--that eliminates the
use of this technology from employers and organizations like,
say, my community health center.
For every dollar we spend on new technology, that is a
dollar that we are taking away from patients, but we do
understand, we have to have a balance in our finances. But in
New York City layering on top and on top of the decision we
make to spend the money on the technology, then to have to now
spend the money on the auditing and all of this.
Essentially, New York City made the decision for us not to
use that technology. That is what happened because it is just
too darned expensive. And I think that there were over one
million nonprofits in the entire country, a million, and many
of these serve underserved communities.
The thing is, this technology can really help us serve
people better. I would just think there needs to be a balance
in whatever standards, rules, laws, regulations are developed
that allow for organizations to use and develop technology
while at the same time--while at the same time protecting
people from bias and from hallucinations that come along with
the technology.
I do not envy your job. That is going to be very, very
difficult. We understand that. But that is where we fall. That
is where----
Senator Hickenlooper. I won't complain. I mean, my
grandfather told me when we were little, if you are lucky
enough to sleep in the bed that you made, don't complain. So, I
won't do that. But I do think it is interesting to get
different viewpoints on where, should the standards come from.
Should it be NIST? Should it be some other maybe even more
global entity that whereby forces people to come together and
make some of the accommodations and assumptions that allow us
to get to a place where, as you are saying, even as we are
continuing to improve those standards, we have got pioneers out
there experimenting and creating new innovations.
Mr. Wilson. But Senator, on the plus side, and I mean on
the really good side is you have people like us that are
willing to help you out with it.
Senator Hickenlooper. Exactly.
Mr. Meyer. We at SHRM, we are fully prepared to help you
out with it.
Senator Hickenlooper. That is why I am asking.
Mr. Meyer. Yes.
Senator Hickenlooper. Mr. Wilson.
Mr. Wilson. Yes, sir. It is a great question, Chairman. Let
me throw you a bone here. So, Seed AI strongly supports the
NIST AI risk management framework, as well as other entities
like the AI Safety Institute and the National AI Research
Resource.
We believe that if these entities were funded and given the
leeway to actually build and test and develop the standards,
that is the direction that we should go for.
Senator Hickenlooper. Youth. I have got lots of youth on
the panel. And Mr. Wilson we didn't get a chance to get to it,
but I thought your story in Houston of how you--the community
college and how you kind of found a whole new avenue through--
why don't you tell a little bit of your story? I don't want to
embarrass you, but you put it in you statement, so.
Mr. Wilson. Yes, it is no problem, sir. Yes, so growing up,
I was more so focused on like football. I am a huge football
guy. I played football 20 years of my life. And I was always
sort of ashamed of showing that I was intelligent.
I was the honors kid in high school, but I had to keep the,
a certain image of, a captain of football team. So once I
graduated, I kind of was able to get more so into what I
actually was passionate about, which is technology and helping
people.
When I got to Houston after moving around through four of
the states, it was a new start for me. And I saw that it was an
environment that allowed people to thrive. It was a healthy
innovation environment.
There were a lot of hungry people. And I don't know if it
was by luck, but when COVID hit and I was laid off and Houston
Community College magically had just started an AI program, I
think it was just a match made in heaven.
I took it and ran with it, and I wanted to galvanize other
students around the same thing. So we built a pretty large
community out there around AI. We actually, me and some of the
students started an AI conference at the school, and the school
ended up picking it up, and they do that event every year now.
There is a lot of good things that came out of that
program. We just kind of want to spread it across the country
and kind of get everybody on the same page as far as what we
need to do.
Senator Hickenlooper. That is something we can all
promote--I love that story. Certainly the potential of AI to
create that catharsis in people I think is tremendous. And
however we work with teachers and teachers unions to make sure
that we accelerate and amplify those opportunities is going to
be powerful.
Whether--I mean College Board, I know that they look at a
different types of AP tests now that might be skills based and
might move in a different direction than what you traditionally
think of as AP chemistry or AP biology. I think AI is going to
allow for that.
We have had several hearings here on apprenticeships and
how important it is to have a lifetime learning. I think almost
all of you said something either in your statements or today
where you referred to that sense that we got to--we are going
to be learning new skills.
Everyone is going to be learning new skills for their whole
lifetime, and how do we facilitate that? In other words, we
need apprenticeships for people of all ages where you can go
and work part time in a different role so that you can acquire
new skills. And that is the, I own--my little secret.
I don't always blurt it out, but we got so excited about
someday having a centralized place with all the skills and
figuring out how do you--how do you have a set of standards
about--if you have got this patch mark, what skills have you
got? How can you demonstrate you are competent with those
skills? But I own and I am donating. This is not a for profit
thing, but I own myshot.com.
If you have ever heard the musical Hamilton where he says,
I am not throwing away my shot, I am not throwing away. Well
that is where myshot.com is going to take all the work that all
of you are doing and have it in one place and allow that
lifetime learning, that application of a lifetime learning to
really, really grow.
Because I think that is the only way we are going to deal
with this. This is going to be a transition over the next 20
years. Look at clean energy. All these things are happening at
once and we are going to look back 50 years, 80 years from now,
going to look back at this as the beginning of the great
transition.
Anyway, you guys are all part of that. Thank you so much
for coming and spending your time. I have got to read my
closing statement, so I don't legally forget something. If
anybody has final statement, something you think the U.S.
Senate should absolutely know? I guess I have worn you out.
In that case, this will end today's hearing. I want to
thank again all of our colleagues who were able to join us in-
person, people watching online, and especially I want to thank
our witnesses, Ms. Kimbrough, Mr. Kotran, Mr. Meyer, Mr.
Wilson. Thank you for your participation. I know you are all
just as busy as we are and you took time out of your lives to
be here for--well, for our Country.
Any Senators who wish to ask additional questions,
questions for the record will be due in 10 business days, so on
October 9th by 5.00 p.m.. The Committee now stands adjourned.
ADDITIONAL MATERIAL
the american federation of labor and congress of industrial
organizations
Thank you Chairman Hickenlooper and Ranking Member Braun for the
opportunity to provide a written statement for the record for your
Subcommittee hearing ``Reading the Room: Preparing Workers for AI.''
The American Federation of Labor and Congress of Industrial
Organizations (AFL-CIO) is a voluntary, democratic federation of 60
affiliated unions representing more than 12.5 million workers in all
sectors of our economy. Our core mission is to ensure that working
people are treated fairly and with respect, that our hard work is
rewarded with family supporting wages and benefits, and that our
workplaces are safe. We also provide an independent voice in politics
and legislation for working women and men and make their voices heard.
We appreciate the Subcommittee's attention to engaging labor unions
and their members in the policy-making process to ensure that workers
shape the future of work. Workers should be protected and positioned to
thrive as artificial intelligence (AI) becomes more widespread in the
workplace. The AFL-CIO looks forward to the HELP Committee building on
this hearing to take a greater role in the work that is needed to
legislate and advance critical policies on AI.
We know that AI systems will have a profound impact on American
society, on the workplace and on the workforce. AI has the potential to
enhance workplace safety and help workers do more meaningful work. It
will also potentially bring great changes to the way many people
perform their work. AI will affect job quality, including technical
standards and best practices, wages and benefits, hiring and retention,
worker rights, safety and well-being, and workforce development.
Therefore, this topic should be of primary importance to the HELP
Committee and its Subcommittees to ensure workers are engaged and
protected from harm and share in the benefits of AI. As the Federal
agency focused on workplaces and workers, the Department of Labor (DOL)
should be heavily engaged in promoting and enforcing AI regulations and
policies.
The AFL-CIO believes that the most beneficial way to prepare the
workforce for AI is to advance a worker-centered approach that ensures
workers are at the hub of AI design, development, implementation, and
the formulation of AI policy that governs its deployment. A worker-
centered AI agenda takes seriously the many ways AI already impacts
workers and gives workers a say over how AI affects workplaces and
jobs. It also ensures AI empowers and benefits workers, safeguards
civil rights and democracy, promotes human-centric creativity, and
guarantees workers access to the training, wraparound support, and
career pathways they need to succeed. Central to this approach is
promoting policies that prioritize and remove legal impediments for
workers to organize, form unions and collectively bargain. Increasing
worker power will positively affect how AI is designed, developed and
implemented and it will ensure that policies related to worker concerns
are addressed: for instance, identifying who is liable if the
technology has negative outcomes or whether there is sufficient
transparency and opportunity for workers to understand and opt out of
data collection.
It is also crucial for policymakers to implement worker-centered,
union-led workforce training so that workers have a direct say in
shaping their future. They must have a voice in AI skills development,
efforts to address job quality and security, safe and ethical AI
implementation, and within meaningful partnerships with employers on
sector-based AI skills strategies.
The AFL-CIO and the AFL-CIO Technology Institute have entered into
an agreement with Microsoft where our affiliated unions and Microsoft
can learn about the real-life, real-time implications of the deployment
of AI. The partnership also includes a neutrality framework for worker
organizing by AFL-CIO affiliate unions. In a recent summit, we were
able to bring together high level decision-makers to have discussions
about the ways that AI is affecting workers on the job and how
technological innovation could be improved by consulting with workers
in the process. Additionally, Microsoft is working with the American
Federation of Teachers, to leverage career and technical education work
that can help prepare today's students for tomorrow's jobs. While we
continue to work to ensure Microsoft and its subsidiaries treat workers
with dignity and respect, we are excited about the possibilities of
partnering with Microsoft to better understand how workers and
corporations can collaborate to create better outcomes in the new
technology development pipeline.
While AI, developed and implemented with workers at the table, may
hold the potential to augment and enhance work, there are examples
today of poorly conceived AI applications that not only harm workers
but also the public at large. A role for worker voice and consent with
regard to AI technologies is necessary to mitigate worker harms, ensure
broad benefit, and safeguard against the use of AI that worsens already
rampant inequality. For example, AI technologies could potentially
increase workplace safety, but could also cause harm by putting workers
under undue stress through work intensification and the override of
human-derived decisions. Workers are on the frontlines of ensuring a
safe, productive workplace and have the firsthand knowledge to inform
how AI is best deployed. It is essential that worker input plays a
significant role in the research and design, development, deployment
and oversight of any AI system. Further, firms should partner with
unions to leverage labor's expertise in developing workforce training
programs, helping to recruit and retain workers, and using workers'
frontline experience to improve the effectiveness and equity of
technology deployment. Partnership with labor is also the best way to
give workers a voice in how AI impacts their work and how the benefits
of increased productivity are shared. We need a cohesive national
approach to workforce development that centers labor-management
partnerships and a national commitment to place-based economic
development with labor at the table.
The AFL-CIO believes it is possible to be both pro-worker and pro-
technology: because of this we believe the most ethical approach to the
adoption and regulation of AI is through policy that requires the
effects on workers be considered, and negative impacts, be mitigated on
the front end. Nonetheless, without proactive policy guardrails,
certain workers are likely to be hard hit by AI and Congress must act
to protect them. A good example exists in the creative industries where
the workers are experiencing unregulated AI, with some devastating
effects, including the ingestion of copyrighted works without
compensation or consent for training AI systems. This threatens the
economic livelihoods of union creative professionals. These individuals
will lose out on collectively bargained royalties, residuals, and
contributions to their health care and retirement funds that come from
the authorized sales and streaming of the creative works they helped
make. Although many of these workers have taken matters into their own
hands by striking for better protections and incorporating protections
into collective bargaining agreements, these and many other examples
across the private and public sector need the attention of Congress--
and specifically this Committee--to ensure that workers are protected.
Hearings that consider the effects of AI on workers are a good
first step to understand the guardrails we need but we must move beyond
thinking that the way to address the issue is to simply prepare the
workforce for the advent of AI. While this is of course an important
part of the adoption of AI, workers have been building, and continue to
build, skills to manage technology advances. To effectively manage AI,
policies must also consider how best to create an environment that will
protect workers from negative consequences like bias, unsafe
workplaces, unfettered workplace surveillance and data collection, and
the removal of worker privacy.
We welcome this hearing, and future engagement with the Committee,
to make further progress toward addressing the potential harms, while
capitalizing on the potential benefits, of AI. The expertise of the
HELP Committee and DOL can be brought to bear on critical AI policy,
design, development and implementation questions. We look forward to
working with the Committee to advance our shared goal of developing
robust AI policy that centers and benefits workers, their families, and
the communities where they live and work.
______
National Retail Federation,
Washington, DC,
September 25, 2024.
Hon. John Hickenlooper, Chairman,
Hon. Mike Braun, Ranking Member,
U.S. Senate HELP Subcommittee on Employment and Workplace Safety,
U.S. Senate Committee on Health, Education, Labor, and Pensions
Washington, DC.
Dear Chairman Hickenlooper, Ranking Member Braun:
On behalf of the National Retail Federation (NRF), I write to thank
you for calling today's hearing entitled, ``Reading the Room: Preparing
Workers for AI.'' I am pleased to share the views of the retail
industry on this important matter. Retailers of all sizes legally and
responsibly utilize automated systems to monitor, manage and evaluate
their workers. NRF respectfully requests that this letter be made a
part of the official record of today's hearing.
The National Retail Federation passionately advocates for the
people, brands, policies and ideas that help retail succeed. From its
headquarters in Washington, DC, NRF empowers the industry that powers
the economy. Retail is the Nation's largest private-sector employer,
contributing $5.3 trillion to annual GDP and supporting more than one
in four U.S. jobs--55 million working Americans. For over a century,
NRF has been a voice for every retailer and every retail job,
educating, inspiring and communicating the powerful impact retail has
on local communities and global economies.
Automated systems have become indispensable tools for the retail
industry, offering benefits including enhanced workplace safety,
productivity, compliance, and security. Retailers have found that such
tools increase organizational efficiency, improve business operations,
contribute to economic growth, and benefit the overall well-being of
their employees.
Retailers' health and safety programs have also been enhanced by
automated systems. For instance, sensors can monitor warehouse
conditions, ensuring that equipment operates safely and temperatures
remain at optimal levels. Wearable technology can reduce physical
strain on employees by providing real-time feedback. In-cabin
monitoring systems can alert truck drivers to fatigue or distractions,
helping to prevent accidents and improve overall safety. Retailers have
found that AI-powered safety platforms can significantly reduce the
occurrence of accidents and unsafe driving incidents, potentially
saving workers' lives.
Productivity monitoring systems can help manage remote workforces,
allocate tasks effectively, and provide objective assessments of
employee performance. These systems can also identify potential
hazards, such as repetitive stress injuries, and help maintain
appropriate staffing levels. By investing in automated systems,
retailers can enhance employee efficiency and create a more productive
work environment.
Automated monitoring systems can play a vital role in ensuring that
workers are paid accurately for their time. By tracking work hours and
minimizing ``off-the-clock'' work, these systems help employers comply
with wage and hour regulations. This not only protects workers' rights
but also reduces the risk of legal disputes and penalties.
Retailers' workplace security has been enhanced by automated
systems. Video surveillance systems can deter theft, violence, and
other security threats, protecting both employees and customers.
Additionally, automated systems can help prevent unauthorized access to
sensitive areas, reducing the risk of theft or loss. By implementing
appropriate security measures, retailers can create a safer and more
secure work environment for their employees.
In all these areas, retailers implement automated systems in
compliance with existing laws and regulations. The current laws on the
books are sufficient to ensure that automated systems protect
employees.
Again, I thank you for calling this hearing and for your attention
to this issue.
Sincerely,
David French,
Executive Vice President,
Government Relations.
______
National Nurses United,
Washington, DC,
September 25, 2024.
Hon. John Hickenlooper, Chairman,
Hon. Mike Braun, Ranking Member,
U.S. Senate HELP Subcommittee on Employment and Workplace Safety,
U.S. Senate Committee on Health, Education, Labor, and Pensions
Washington, DC.
Dear Chairman Hickenlooper, Ranking Member Braun, and Members of
the Committee:
In light of the Committee's hearing today titled ``Reading the
Room: Preparing Workers for AI,'' I write to you on behalf of National
Nurses United, the Nation's largest union and professional association
of registered nurses (RNs). Our nearly 225,000 members primarily work
in acute care hospitals, where they are already working with and
experiencing the impacts of artificial intelligence (AI) and data-
driven technologies at the hospital bedside. Given the lack of health
care worker representation at the hearing, we wanted to share directly
with you why it's essential that health care workers have the right to
be consulted and engaged on all policies, procedures, and best
practices surrounding the implementation of AI.
The health care industry has been implementing various forms of
artificial intelligence and other data driven technologies for a number
of years. The nursing workforce is therefore uniquely situated to
provide feedback and analysis on the impacts that these technologies
have had on workers and on patients. In the experience of nurses, the
decisions to implement these technologies are often made without the
knowledge of either nurses or patients and are putting patients and the
nurses who care for them at risk. Technologies that have already been
implemented include the clinical decision support systems embedded in
electronic health records (EHRs), acute-care hospital-at-home and
remote patient monitoring schemes, virtual acute-care nursing,
automated worker surveillance and management (AWSM), staffing platforms
that support gig nursing and understaff hospital units to unsafe
levels, and increasingly, emerging technologies like generative AI
systems.
Many of these technologies may be used to track the activities of
health care workers and interfere in union activity. Automated
monitoring technology feeds into algorithmic management systems that
make unreasonable and inaccurate decisions about patient acuity,
staffing, and care with the goal of lowering labor costs. As a result,
nurses and other health care professionals are expected to work faster,
accept more patients per nurse than is safe, and reduce nurses' use of
independent professional skill and judgment, ultimately endangering
patients. Studies show that when RNs are forced to care for too many
patients at one time, patients are at higher risk of preventable
medical errors, avoidable complications, falls and injuries, \1\
pressure ulcers, \2\ increased length of hospital stay, higher numbers
of hospital readmissions, and death. \3\
---------------------------------------------------------------------------
\1\ Kim J, Lee E, Jung Y, Kwon H, Lee S. Patient-level and
organizational-level factors influencing in-hospital falls. J Adv Nurs.
2022 Nov;78(11):3641-3651. doi: 10.1111/jan.15254. Epub 2022 Apr 20.
PMID: 35441709; PMCID: PMC9790490.
\2\ Kim J, Lee JY, Lee E. Risk factors for newly acquired pressure
ulcer and the impact of nurse staffing on pressure ulcer incidence. J
Nurs Manag. 2022 Jul;30(5):O1-O9. doi: 10.1111/jonm.12928. Epub 2020
Feb 25. PMID: 31811735; PMCID: PMC9545092.
\3\ Increased LOS, Mortality and Readmission: Dierkes, A. M.,
Aiken, L. H., Sloane, D. M., Cimiotti, J. P., Riman, K. A., and McHugh,
M. D. (2022). Hospital nurse staffing and sepsis protocol compliance
and outcomes among patients with sepsis in the USA: a multistate cross-
sectional analysis. BMJ Open, 12(3), e056802. https://doi.org/10.1136/
bmjopen-2021-056802.
Additionally, tracking nurse actions is designed to facilitate
routinization--breaking the holistic process of nursing into discrete
tasks--with the goal of replacing educated registered nurses exercising
independent judgment with unlicensed or lower-licensed staff following
algorithmic instructions to reduce labor costs. Registered nurses have
extensive education and clinical experience that enables them to
provide safe, therapeutic patient care. Even the simplest RN-patient
interactions involve skilled assessment and evaluation of the patient's
overall condition. Subtle changes in a patient's skin tone, respiratory
rate, demeanor, and affect provide critical information about their
health and well-being that can be easily overlooked or misinterpreted
by those without an RN's education and clinical experience. Attempts to
routinize and deskill registered nursing practice undermines safe
---------------------------------------------------------------------------
patient care.
Employers generally assert that these powerful technologies are
just updates of older technology that has long been in the workplace,
such as treating computer-vision aided cameras the same as traditional
security cameras, or EHRs as electronic versions of old paper medical
records. However, these technologies are much more than modern
iterations of well understood tools and are being introduced widely
despite lack of robust research showing safety, reliability,
effectiveness, and equity. Rather, AWSM technologies pull vast and
diverse data from an entire ecosystem of monitoring equipment and
process this information through opaque algorithms that then make
clinical and employment decisions. There is no current method for
evaluating AI being used in nursing practice applications, and no
requirement for external validation; it is clear to nurses that AI
technologies are being designed to be a replacement for skilled
clinicians as opposed to a tool that many clinicians would find
helpful.
The right for union workers to bargain over whether and how
technology should be implemented in the workplace before it is selected
or deployed must be protected. Workers have the right to understand how
the decisions governing their working lives are made--including
hiring--and patients have the right to understand how the decisions
concerning their care are made, including insurance coverage
determinations. Through our experiences working with and around these
systems, it is clear to registered nurses that hospital employers have
used these technologies in attempts to outsource, devalue, deskill, and
automate our work. Doing so increases their profit margins at the
expense of patient care and safety.
In our experience in the health care sector, labor-management
partnerships do not serve the interests of patients or the workers who
care for them, instead benefiting the employer. Proposals to utilize
labor management partnership to self-regulate AI in the healthcare
setting are insufficient and would be ineffective. They must not be
used in place of comprehensive pre-market testing, approval and
monitoring by a regulatory agency. Congress must both ensure strong
regulation of AI, while also strengthening the rights of workers to
organize and bargain collectively.
Nurses have grave concerns about the fundamental limits on the
ability of algorithms to meet the needs of individual patients,
especially when those patients are part of racial or ethnic groups that
are less well represented in the data. Nurses know that clinical
algorithms can interfere with safe, therapeutic health care that meets
the needs of each individual patient. While clinical algorithms may
purport to be an objective analysis of the scientific evidence, in fact
their development involves significant use of judgment by their
creators and creates the opportunity for creator bias--from conflicts
of interest, limited perspective on the lives of racial minorities, or
implicit racial bias--to be introduced into the algorithm.
In sum, we are deeply concerned about the use of AI in nursing
practice in acute care hospitals. We appreciate the subcommittee's
interest in examining the impacts of artificial intelligence (AI) and
data-driven technologies on the U.S. workforce. As the Nation's largest
union of registered nurses, we look forward to working with you on this
important matter.
Sincerely,
Amirah Sequeira,
National Government Relations Director,
National Nurses United.
______
QUESTIONS AND ANSWERS
Response by Dr. Karin Kimbrough to Questions of Senator Lujan
senator lujan
Question 1. In your written testimony, you break down the labor
impacts on AI and photography is given as an example of a profession
that will be disrupted. How does your research team at LinkedIn foresee
AI impacting the work of creatives (including visual artists, authors,
photographers, filmmakers, screenwriters, musicians)?
Answer 1. All tech evolutions start with disruption, but in the
aggregate the impact is likely to be net positive because for the vast
majority of people, AI isn't replacing their job but transforming it,
and their next job might even be a role that doesn't exist yet.
Globally, when we look at our data, skills for the average job are
projected to change by 50 percent by 2030 compared to 2016--and
generative AI is expected to accelerate this change rate to nearly 70
percent.
While we don't discount the challenges disruptive technologies can
create for workers, we do nonetheless expect to see job creation as a
result of AI. History tells us that new waves of technological
innovation tend to create more jobs than they displace. Just as web
developers, cybersecurity experts, social media managers, and content
creators proliferated with the rise of the internet, new AI roles--
across ethics, research, design and other fields--will emerge in the
coming years.
With respect to the creative fields, it is, in fact, likely that
jobs will be disrupted by AI. However, we don't see disruption as
eliminating these jobs completely, but rather changing how jobs are
done and the skills needed for these occupations. For example, authors
may use AI to help them generate ideas, develop plot structures, and
supplement background research for their work. Academic studies have
shown that AI is useful in jumpstarting projects, but can struggle when
generating truly novel ideas and struggle with tasks that are beyond
the frontier of their training data.
Question 2. What can be done to ensure that creative industry
workers whose work is disrupted by generative AI are not driven from
their profession entirely?
Answer 2. At a high level, it is important for workers to
understand both how their job will be impacted by AI (i.e. how the
skills or technology they use will change) and what training is
available to them to adapt and learn new skills that allow them to
leverage AI and augment their work, rather than eliminating what they
are doing. We also think that developers and deployers of AI tools
should be thinking about this impact and incorporating worker voice in
ways that not only improve the products and services being developed,
but are also mindful of the impact they will have on workers.
Additionally, LinkedIn Learning offers a variety of courses to help
creative professionals leverage AI in their jobs. From general courses
like ``Integrating Generative AI into the Creative Process'' to courses
aimed at specific creative professions, like ``Photoshop and Lightroom:
Portrait Retouching in the Age of AI'' for photographers.
______
Response by Mr. Wilson to Questions of Senator Lujan
senator lujan
Question 1. What impact does broadband access and affordability--
and the current digital divide that exists in the U.S.--have on the
efforts to accelerate AI preparedness for teachers, students and future
workers?
Answer 1. Access to the internet is an essential requirement for
any person--student, teacher, current and future workers, or
otherwise--to meaningfully participate in an AI-driven economy. If a
person cannot afford or otherwise access broadband, their ability to
learn about and use AI, co-create with the technology, and seek new
opportunities to gain and leverage AI skills will be severely limited.
Question 2. In your written testimony, you discuss the importance
of building accessible and affordable AI infrastructure, including
data, compute, models and training. In addition to these components of
AI infrastructure, does Seed AI agree that high speed broadband access
and affordability is a critical component that is necessary for
enabling AI preparedness?
Answer 2. Yes, SeedAI agrees that affordable and accessible
Internet access is a necessary part of a broader strategy to promote AI
preparedness.
______
[Whereupon, at 3:25 p.m., the hearing was adjourned.]
[all]