[House Hearing, 119 Congress]
[From the U.S. Government Publishing Office]


                    FROM CHALKBOARDS TO CHATBOTS: THE
                     IMPACT OF AI ON K	12 EDUCATION

=======================================================================

                                HEARING

                               Before The

                   SUBCOMMITTEE ON EARLY CHILDHOOD, 
                  ELEMENTARY, AND SECONDARY EDUCATION

                                 OF THE

                  COMMITTEE ON EDUCATION AND WORKFORCE
                     U.S. HOUSE OF REPRESENTATIVES

                             FIRST SESSION

                               __________


             HEARING HELD IN WASHINGTON, DC, APRIL 1, 2025

                               __________

                            Serial No. 119-6

                               __________

    Printed for the use of the Committee on Education and Workforce
    
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]    


        Available via: edworkforce.house.gov or www.govinfo.gov
        
                                __________

                   U.S. GOVERNMENT PUBLISHING OFFICE                    
61-537 PDF                  WASHINGTON : 2025                  
          
-----------------------------------------------------------------------------------     
       
                  COMMITTEE ON EDUCATION AND WORKFORCE

                    TIM WALBERG, Michigan, Chairman

JOE WILSON, South Carolina           ROBERT C. ``BOBBY'' SCOTT, 
VIRGINIA FOXX, North Carolina            Virginia,
GLENN THOMPSON, Pennsylvania           Ranking Member
GLENN GROTHMAN, Wisconsin            JOE COURTNEY, Connecticut
ELISE M. STEFANIK, New York          FREDERICA S. WILSON, Florida
RICK W. ALLEN, Georgia               SUZANNE BONAMICI, Oregon
JAMES COMER, Kentucky                MARK TAKANO, California
BURGESS OWENS, Utah                  ALMA S. ADAMS, North Carolina
LISA C. McCLAIN, Michigan            MARK DeSAULNIER, California
MARY E. MILLER, Illinois             DONALD NORCROSS, New Jersey
JULIA LETLOW, Louisiana              LUCY McBATH, Georgia
KEVIN KILEY, California              JAHANA HAYES, Connecticut
MICHAEL A. RULLI, Ohio               ILHAN OMAR, Minnesota
JAMES C. MOYLAN, Guam                HALEY M. STEVENS, Michigan
ROBERT F. ONDER, Jr., Missouri       GREG CASAR, Texas
RYAN MACKENZIE, Pennsylvania         SUMMER L. LEE, Pennsylvania
MICHAEL BAUMGARTNER, Washington      JOHN W. MANNION, New York
MARK HARRIS, North Carolina          VACANCY
MARK B. MESSMER, Indiana
VACANCY

                     R.J. Laukitis, Staff Director
              Veronique Pluviose, Minority Staff Director
                                 ------                                

  SUBCOMMITTEE ON EARLY CHILDHOOD, ELEMENTARY, AND SECONDARY EDUCATION

                   KEVIN KILEY, California, Chairman

MARY E. MILLER, Illinois             SUZANNE BONAMICI, Oregon,
GLENN THOMPSON, Pennsylvania           Ranking Member
BURGESS OWENS, Utah                  JAHANA HAYES, Connecticut
MICHAEL A. RULLI, Ohio               SUMMER L. LEE, Pennsylvania
JAMES C. MOYLAN, Guam                JOHN W. MANNION, New York
RYAN MACKENZIE, Pennsylvania         FREDERICA S. WILSON, Florida
MARK HARRIS, North Carolina          ALMA S. ADAMS, North Carolina
MARK B. MESSMER, Indiana
                        
                        
                        C  O  N  T  E  N  T  S

                              ----------                              
                                                                   Page

Hearing held on April 1, 2025....................................     1

                           OPENING STATEMENTS

    Kiley, Hon. Kevin, Chairman, Subcommittee on Early Childhood, 
      Elementary, and Secondary Education........................     1
        Prepared statement of....................................     4
    Bonamici, Hon. Suzanne, Ranking Member, Subcommittee on Early 
      Childhood, Elementary, and Secondary Education.............     7
        Prepared statement of....................................    10

                               WITNESSES

    Dobrin, Dr. Sid, Chair, Department of English, University of 
      Florida....................................................    13
        Prepared statement of....................................    15
    Rafal-Baer, Dr. Julia, CEO, ILO Group........................    20
        Prepared statement of....................................    22
    Mote, Erin, CEO, InnovateEDU/EDSAFE AI Alliance..............    25
        Prepared statement of....................................    28
    Chism, Chris, Superintendent, Pearl Public School District...    43
        Prepared statement of....................................    45

                        QUESTIONS FOR THE RECORD

    Responses to questions submitted for the record by:
        Ms. Erin Mote............................................    84

 
                   FROM CHALKBOARDS TO CHATBOTS: THE
                     IMPACT OF AI ON K-12 EDUCATION

                              ----------                              


                         Tuesday, April 1, 2025

                  House of Representatives,
  Subcommittee on Early Childhood, Elementary, and 
                               Secondary Education,
                      Committee on Education and Workforce,
                                                    Washington, DC.
    The Subcommittee met, pursuant to notice, at 10:16 a.m., in 
Room 2175, Rayburn House Office Building, Hon. Kevin Kiley 
(Chairman of the Subcommittee) presiding.
    Present: Representatives Kiley, Miller, Thompson, Owens, 
Rulli, Moylan, Mackenzie, Harris, Messmer, Bonamici, Hayes, 
Lee, Mannion, Wilson, Adams, and Scott.
    Staff present: Vlad Cerga, Director of Information 
Technology; Maren Emmerson, Intern; Dara Gardner, Einstein 
Fellow; Amy Raaf Jones, Director of Education and Human 
Services Policy; Libby Kearns, Press Assistant; Campbell Ladd, 
Clerk; R.J. Laukitis, Staff Director; Danny Marca, Director of 
Information Technology; R.J. Martin, Professional Staff Member; 
Audra McGeorge, Communications Director; Eli Mitchell, 
Legislative Assistant; Ethan Pann, Deputy Press Secretary and 
Digital Director; Kane Riddell, Staff Assistant; Sara 
Robertson, Press Secretary; Brad Thomas, Deputy Director of 
Education and Human Services Policy; Ann Vogel, Director of 
Operations; Ali Watson, Director of Member Services; James 
Whittaker, General Counsel; Ellie Berenson, Minority Press 
Assistant; Bryan Gonzalez, Minority Grad Intern; Rashage Green, 
Minority Director of Education Policy & Counsel; Christian 
Haines, Minority General Counsel; Emanual Kimble, Minority 
Professional Staff; Raiyana Malone, Minority Press Secretary; 
Ben Noenickx, Minority Intern; Eleazar Padilla, Minority Staff 
Assistant; Veronique Pluviose, Minority Staff Director.
    Chairman Kiley. The Subcommittee on Early Childhood, 
Elementary, and Secondary Education will come to order. I note 
that a quorum is present. Without objection, the Chair is 
authorized to call a recess at any time.
    Good morning. Artificial intelligence has been advancing at 
such a rapid pace in recent months, weeks, and days, that by 
the end of this hearing anything we say this morning will 
probably be outdated. That is perhaps, a slight exaggeration, 
but it is the essential reality. Leading labs continue to put 
out new models each week, shattering benchmarks, demonstrating 
incredible capabilities and pointing toward still greater 
advancements ahead.
    Indeed, there does not appear to be any limit to this 
progress in sight. This rapid acceleration toward the future 
has brought a host of anxieties, not least of which is 
geopolitical. At this moment America holds a clear, but 
precarious AI advantage. There are also concerns related to 
jobs, privacy, safety, control and more broadly, a sense of 
uncertainty about the social changes that come and break 
through the wall herald, and what our world will look like for 
the next generation.
    As understandable and important as these concerns are, the 
bigger picture of one of opportunity, and a truly limitless 
sense of possibility. Even with the state-of-the-art, as it now 
exists, let alone what it will be next week, next year, or a 
decade from now, we suddenly have tools to address many long-
standing challenges in new and powerful ways.
    Education is one very clear example of that. I am a former 
high school teacher, and I believe one of the greatest failings 
in our country's modern history is the way millions of kids 
have been deprived of a decent education. Our school system has 
shameful achievement gaps, nowhere more so than in my home 
State of California.
    Open to the neighborhood they were born into, far too many 
young people in this country are not receiving the education 
they deserve and are robbed of the opportunities that a quality 
education provides. Meanwhile, the educational attainment of 
the country has a whole has been on a sharp decline.
    AI can change that. It has the potential to give every 
child in America a richer educational experience than any child 
in America had just a few years ago. It can give every teacher 
in America a greater ability to reach students than any teacher 
did in the past.
    It can empower parents to follow their child's progress and 
guide it accordingly. For a glimpse of this potential, here is 
a short video.
    [Video played.]
    At the most basic level, AI tools like Khanmigo from the 
Khan Academy give every child access to a world class tutor in 
any subject. This is no small matter. Studies have shown that 
just a hear of high dosage tutoring can add up to a full 2 
years worth of additional learning gains. This is a very 
special kind of tutor, one that is available any time of the 
day for unlimited amounts of time, and that cannot be stumped 
by any question.
    It is a tutor that is thoroughly familiar with the 
student's strengths and weaknesses and baseline knowledge, 
enabling personalized instruction that compels critical 
thinking without giving away answers. Perhaps best of all, this 
tutor is low-cost or free.
    For most of history, access to this kind of personalized 
instruction would have been completely unthinkable, yet, AI has 
made it commonplace. These tools can customize the learning 
experience for each student, adapting their content, pace and 
learning style to the student's performance and preferences.
    They can also bring learning to life in new and dynamic 
ways. Learning physics from Albert Einstein or engaging with a 
fictional character when studying work of literature. We are 
already seeing AI widely adopted by students in limited ways. 
Nearly 50 percent K12 students use ChatGPT at least weekly.
    Of the students who use AI, 35 percent use it to summarize 
information, 32 percent to generate ideas for assignments, and 
26 percent to get initial feedback on their work. The question 
is not whether students will use AI, that is already happening. 
Rather, the question is how schools can support students in 
using AI responsibly and in unlocking its full potential to 
advance student achievement.
    It is not just students who stand to benefit. AI can 
empower teachers and school leaders to fulfill their vision for 
their classrooms, and to connect with students in the highest 
impact ways. AI tools can help with tasks like lesson planning 
and grading and can free teachers to focus on the aspects of 
education that only a caring human can provide.
    A recent McKinsey analysis found that AI can save teachers 
up to 13 hours per week. There are also, of course, risks when 
it comes to AI usage in the classroom. Recent studies show the 
ready availability of Chatbots has proven an irresistible 
temptation to many students.
    Nearly 40 percent of middle and high school students 
admitted they used AI without teacher's permission to complete 
assignments, according to a survey last year. Now, while 
outright cheating is certainly a concern, the bigger challenge 
is to be vigilant in assuring AI never becomes a shortcut to 
avoid engaging in critical thinking, formulating original ideas 
and persevering through challenging content.
    Student privacy is also a crucial issue. District leaders 
must be thoughtful about the data that gets collected and 
vigilant about securing that data. We will hear testimony today 
about how district leaders and administrators can use AI 
responsibility without endangering students' rights.
    Finally, we must be mindful of excessively absorbing 
students into digital worlds. The negative effects of smart 
phone use for you people have become impossible to deny. The 
last thing we want to do is compound this problem. I suspect 
the most successful approaches to AI in the classroom will 
assure human interaction remains fundamental to the educational 
experience.
    None of this, I should add, is a call for new mandates at 
the Federal level. Quite the contrary, education is 
fundamentally a State and local issue, and the best education 
solutions emerge when school leaders and teachers are given 
flexibility to do what is best for their students.
    This is especially true given the infinite variety of ways 
AI can be utilized in the classroom, and the rapid pace at 
which it continues to evolve. The purpose of today's hearing is 
to cast light on the enormous potential that exists. We can 
continue to highlight outstanding examples of AI in education 
throughout the country, so best practices can be shared.
    We have an excellent witness panel assembled to give us 
perspective on all of these questions, and I am looking forward 
very much to hearing their thoughts. With that, I will yield to 
the Ranking Member for an opening statement.
    [The statement of Chairman Kiley follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]    

    Ms. Bonamici. Thank you, Chairman Kiley, and thank you to 
the witnesses for being here today. I am encouraged that all of 
the witnesses submitted testimony that recognizes the 
opportunities and the challenges of artificial intelligence in 
K12 education.
    We are at a pivotal time in education, and the path we 
choose will determine whether we can equip students and 
educators with the tools, the critical thinking skills, and the 
knowledge they need to succeed in an ever evolving, 
technologically driven world.
    Artificial intelligence is not in the distant future; it is 
in our schools right now. It is shaping how students are 
educated, and how they learn. Although this is an exciting 
prospect, it also presents significant challenges, particularly 
regarding equity and regulation.
    Without a doubt, dismantling the Department of Education 
will exacerbate inequities, and set us back as a Nation. The 
Department of Education's not just a building. Through its 
programs it plays a pivotal role in closing achievement gaps, 
and helping to meet the goals that all students, including 
those in rural and low-income communities, and including 
students with disabilities, have an opportunity to access high-
quality public education.
    It is the Department of Education that helps level the 
playing field and provides critical resources that many 
districts are not able to afford without Federal funding. Title 
I for example, provides resources for schools with high 
concentration of poverty, a lifeline for the schools that need 
it the most. In fact, every witness here today mentioned in 
their testimony the importance of access, which is one of the 
main points of Title I.
    It is unclear what will happen to Title I, but without 
experienced implementation at the Department of Education, it 
is likely, assuming it survives, we will see support severely 
cut or limited, and the prospect of waivers to block grant 
Title I, which some republicans are advocating for is troubling 
because without accountability systems we do not know if the 
dollars will go to the highest need schools, resulting in 
deepening inequities and academic gaps.
    It is also important to mention that the Trump 
administration cut funding for a significant portion of Federal 
grants that support education, educator professional 
development, including the teacher quality partnership. 
Thankfully, Federal Courts have ordered the administration to 
restore these funds.
    Professional development opportunities are crucial to equip 
educators with evidence-based teaching practices, as well as 
prepare educators on AI technologies and other necessary skills 
needed to educate students for the 21st Century.
    It is no question that a loss of funding, or inconsistent 
funding, or delayed funding disbursement, will 
disproportionately harm struggling low-income and rural schools 
and students. Without a robust and equitable funding system 
with a strong accountability framework the digital divide will 
widen.
    Technology disparities will mirror existing inequalities, 
and that will leave vulnerable students with fewer 
opportunities to benefit from AI integrated learning 
environments, or to learn proper guardrails surrounding the use 
of AI. During the 1990's and early 2000's as technology was 
rapidly evolving, it was the Department of Education that led 
efforts to close the digital divide.
    Without Federal leadership we would have seen even greater 
inequities. Unfortunately, today the Department of Education is 
still reeling from a significant reduction in force, including 
the elimination of the Office of Educational Technology. This 
office was instrumental in guiding schools on the safe and 
ethical integration of new technologies like AI, providing 
resources on data security and best practices.
    With the OET gone, we risk losing schools--leaving schools 
and students unprepared for AI's opportunities and challenges. 
Related to the topic today, the massive chainsaw cuts to IES, 
the Institute of Education Sciences is absurd. Research helps 
educators and policymakers make good decisions, and that is not 
something local school districts and states can easily 
replicate.
    Several states, including Virginia, California and my home 
State of Oregon, have started developing guidance and policies 
about AI use in the classroom. Without the Department of 
Education's leadership, states are left to navigate this 
complex landscape on their own, which again creates 
inconsistencies and exacerbates achievement gaps.
    We recognize that local schools and districts have the 
authority already to determine local decisions, whether it be 
teacher standards, class size, curriculum, that is local. It is 
this Federal investment in research and leadership that makes a 
difference.
    As we grapple with the role of AI in education, we must 
equip students with the skills necessary to survive in today's 
society and economy. The jobs of tomorrow will demand 
proficiency in technology, including in AI.
    Without proper education of students, especially those from 
underfunded districts, will find themselves at a disadvantage, 
unable to compete in an increasingly globalized economy. This 
issue is not just about access to technology. It is how we use 
the technology.
    As AI is integrated into classrooms, we must be diligent, 
so that it does not reinforce existing biases, or create new 
ones. With the Department staffing cuts and an overwhelming 
caseload, The Office for Civil Rights in the Department of 
Education, which addresses discrimination in schools through 
investigation.
    They are struggling to keep up. We need Federal leadership 
and research to help guide the use of AI in schools, so all 
students, regardless of their ZIP Code, income, race, have the 
opportunity to learn, grow and thrive in this new digital age.
    That is clear from the testimony of the witnesses today, 
and it is also clear that eliminating the Department of 
Education, and for that matter, the National Science 
Foundation, and other research-based entities is antithetical 
to that goal.
    As we move forward, Democrats are committed to providing 
equal access to public education for all students. We are 
looking for reasonable policymakers from either side of the 
aisle to join us in strengthening local schools by preserving 
these important Federal investments in education and education 
research.
    This should not be a partisan issue. It affects every 
student, teacher and family in the country, and it most 
certainly affects our potential for growth in competitiveness 
in a global economy. I look forward to the conversation, and I 
yield back the balance of my time.
    [The statement of Ranking Member Bonamici follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]    

    Chairman Kiley. Pursuant to Committee Rule 8-C, all members 
who wish to insert written statements into the record may do so 
by submitting them to the Committee Clerk electronically in 
Microsoft Word format by 5 p.m., 15 days after this hearing. 
Without objection, the hearing record will remain open for 14 
days to allow such statements and other extraneous material 
noted during the hearing to be submitted to the official 
hearing record.
    I note for the Subcommittee that some of my colleagues who 
are not permanent members of the Subcommittee may be waving on 
for the purpose of today's hearing. I will now introduce our 
distinguished witnesses.
    Our first witness is Dr. Sid Dobrin, the Chair of the 
Department of English at the University of Florida in 
Gainesville, Florida. Our second witness is Dr. Julia Rafal-
Baer, the CEO of the ILO Group here in Washington, DC. Our 
third witness is Ms. Erin Mote, the CEO, for InnovateEDU in 
Brooklyn, New York.
    Our fourth witness is Mr. Chris Chism, the Superintendent 
of the Pearl Public School District in Pearl, Mississippi. We 
thank the witnesses for being here today, and we look forward 
to your testimony.
    Pursuant to Committee rules, I would ask that you each 
limit your oral presentation to a 3-minute summary of your 
written statement. The clock will count down from 3 minutes as 
Committee members have many questions for you, and we would 
like to spend as much time as possible on those questions.
    However, pursuant to Committee Rule 8D and Committee 
practice, we will not cutoff your testimony until you reach the 
5-minute mark. I would like to remind the witnesses to be aware 
of their responsibility to provide accurate information to the 
Subcommittee, and I will first recognize Dr. Dobrin for your 
testimony.

  STATEMENT OF DR. SID DOBRIN, CHAIR, DEPARTMENT OF ENGLISH, 
                     UNIVERSITY OF FLORIDA

    Mr. Dobrin. We were not given technical instructions. Mr. 
Chairman, Ranking Committee Member, and distinguished members 
of the Committee and others in attendance, I want to thank you, 
first, for the opportunity to speak with you today about 
artificial intelligence and generative artificial intelligence 
in education specifically within K through 12 education.
    As noted, my name is Sid Dobrin. I am Chair of the 
Department of English at the University of Florida where I have 
been faculty for 28 years now. My primary focus of research 
falls to writing studies specifically in terms of emerging 
technologies such as augmented reality, virtual reality, and 
artificial intelligence.
    I will not bore you with my other bonafides, since you have 
them in front of you, other than to say over the last 2 years I 
have had the opportunity to speak to more than 70 campuses 
worldwide about integrating artificial intelligence into 
curriculum.
    I have to tell you, I very much appreciate getting those 
invitations, and the invitation to speak with you today because 
it usually goes something like this. Hey Sid, could you come 
talk to us about artificial intelligence and generative 
artificial intelligence in education in these massive, galactic 
subjects, and if you can do us a favor and wrap it up in 3 
minutes.
    These are big, complex subjects, no matter what I say today 
we'll only be scratching the surface of the conversations we 
need to be having. Since the subject of, and research about 
artificial intelligence and generative artificial intelligence 
in the workplace is such a complex and extensive conversation, 
what I am going to talk about now really is just sort of 
introductory.
    The introduction of generative AI tools like ChatGPT have 
marked a turning point in education. While many became aware of 
these technologies in November 2022, the groundwork for AI 
education was laid much earlier. Initiatives such as the 
Artificial Intelligence for K12 Initiative, developed by 
organizations like the Association for the Advancement of 
Artificial Intelligence, and the Computer Science Teacher 
Association, aimed to create national guidelines and resources 
for AI education.
    However, the rapid evolution of AI necessitates a broader 
approach beyond computer science alone. Today, integrating AI 
into K through 12 education is crucial for several reasons. 
First, technological advancements have made powerful AI tools 
widely accessible, and as was noted in the opening statements, 
access is a key conversation that we must be having.
    It has also changed how students learn to interact with 
information. It is not enough for students to merely consume 
these technologies, they must also understand how they work. 
This comprehension will prepare them for the modern workplace 
where AI is becoming increasingly integral.
    Incorporating AI literacy into curriculum fosters critical 
thinking and problem-solving skills essential for success in 
any career path. Employers are increasingly looking for 
candidates who can navigate complex AI tools, analyze data, and 
make informed decisions.
    By embedding AI education across subjects, we can equip 
students with these vital competencies, building a more 
adaptable, innovative workforce ready to meet the demands of a 
rapidly changing job market. Moreover, as we transition into an 
AI driven economy, workforce readiness will hinge on our 
educational systems' ability to prepare students for these new 
realities.
    We must ensure that our students are not just passive users 
of technology, but proactive participants who can leverage AI 
to enhance their creativity and productivity. Thus, in my 
comments today, and in my written statement, I urge the 
Committee to prioritize AI integration into K through 12 
education to prepare our students for the future.
    By doing so, we not only enhance individual career 
prospects, but also contribute to American innovation and 
economic competitiveness on a global scale, which I address in 
my written testimony. I want to thank you for your time, and 
for the opportunity to speak with you about this pressing 
issue. Thank you.
    [The statement of Dr. Dobrin follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    
    Chairman Kiley. Thanks very much. We shall now recognize 
Dr. Rafal-Baer for your testimony.

STATEMENT OF DR. JULIA RAFAL-BAER, CEO, ILO GROUP, WASHINGTON, 
                              D.C.

    Ms. Rafal-Baer. Thank you. Chair Kiley, Ranking Member 
Bonamici, and members of the Committee, thank you for the 
opportunity to testify today regarding the integration of 
artificial intelligence in K12 education.
    My name is Dr. Julia Rafal-Baer, I am the CEO of ILO Group. 
We are a women owned, women led policy and strategy firm. We 
work with over 200 districts and states nationwide.
    Our work impacts one in three students in the country. I am 
also the founder of Women Leading Ed. We are the largest 
national nonprofit network for women in educational leadership. 
We have members in every State in this Nation. Today, I advise 
education leaders nationwide.
    I serve on multiple Federal and national boards, including 
the National Assessment Governing Board. AI is rapidly becoming 
one of the most impactful innovations across all industries. It 
has the power to transform how educators teach, how students 
learn, how schools operate, and how prepared our future 
workforce will be to leverage emerging technologies.
    However, Women Leading Ed did a survey of hundreds of women 
across the country, and what we learned is that 81 percent of 
leaders feel that AI is significantly impacting their systems. 
However, fewer than 10 percent report that they have the 
adequate resources for effective and safe implementation.
    Given this reality, we must ensure that every student, 
educator, caregiver and community member is equipped not merely 
with digital literacy, but with AI literacy.
    The 2024 bipartisan House Task Force on AI rightly 
recognizes the importance of AI literacy, and the need for 
resources for it. Our work has also found that to establish 
robust foundations for responsible AI usage there must be 
meaningful and ongoing community engagement.
    Based on ILO Group's specific work on AI for the past two 
and a half years, I offer you three specific recommendations 
for Federal policy. First, the Federal role should be 
intentionally limited. It should allow states and districts the 
flexibility and the autonomy to lead AI integration efforts 
that reflect their unique context.
    Given the rapidly evolving role and the nature of these 
technologies, decisions about AI literacy specifically, should 
remain local, rather than federally defined. Second, 
strengthening data security and privacy protections must be an 
immediate priority as AI adoption accelerates.
    Here the Federal Government has an appropriate role to 
provide consistent cybersecurity and data guidance across all 
states, and across all agencies. Education systems increasingly 
face sophisticated AI enhanced cybersecurity threats, making 
standardized protections essential.
    Federal policymakers could develop a comprehensive data 
privacy bill that supports effective policies and best 
practices related to online safety and data privacy that are 
broadly disseminated to states.
    Third, the Federal Government has a critical role in 
funding research that rigorously assesses AI's impact on 
education. The Federal Government is uniquely positioned to 
research and evaluate where rapidly evolving AI tools enhance 
educational quality, relevance and efficiency, and also to 
evaluate its risks, and conduct cost benefit analyses at scale.
    Federal support is essential for analyzing and sharing 
evidence-based practices and ensuring that educator and student 
privacy is adhered to. We recommend convening a White House 
Summit on AI and Education to examine both the opportunities 
and these limitations.
    In conclusion, thoroughly integrated AI can transform 
educational outcomes, but achieving this is going to require 
careful planning, prioritizing deep stakeholder engagement, 
comprehensive AI literacy, and robust security measures. Thank 
you again for your leadership on this critical issue, and for 
the opportunity to speak with you today.
    [The statement of Dr. Rafal-Baer follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]    

    Chairman Kiley. I will now recognize Ms. Mote for your 
testimony.

    STATEMENT OF MS. ERIN MOTE, CEO, INNOVATEEDU/EDSAFE AI 
                  ALLIANCE, BROOKLYN, NEW YORK

    Ms. Mote. Good morning, Chairman Kiley, Ranking Member 
Bonamici, and members of the Committee. Thank you for the time 
and attention today. I appreciate the chance to address you 
about the State of AI in K12 education. My remarks are shaped 
by decades of work at the intersection of education and 
technology.
    As a technologist, and enterprise architect, I have been 
using AI and engineering technology for schools, including the 
charter school that I founded in downtown Brooklyn for more 
than a decade. My comments today are informed by my experience 
working alongside State Chiefs, District Administrators, 
educators, families, communities, through the EDSAFE AI 
Alliance and InnovateEDU.
    The Alliance fosters collaboration and knowledge to 
prioritize student well-being, and effective learning outcomes 
through the SAFE framework, safety, accountability, fairness 
and transparency, and efficacious use of AI in education. 
Though, my most important role is that of a mother of two 
school-aged children.
    Like some of you, and all parents across this country, I 
want to ensure that my children are safe, that they are known 
and cared for, that their data is safe, their teachers are 
supported, and they have a pathway to join a future workforce 
where a pre-requisite for any career, from a lawyer to doctor, 
to plumber, will be of how you work alongside, and manage 
artificial intelligence.
    Artificial intelligence is an arrival technology. Arrival 
technologies, like electricity, social media, or the internet, 
are massively disruptive, and rapidly reshape how we teach, how 
we learn, how we live. Right now, AI's use is uneven and 
limited across school communities, with only about 25 percent 
of educators reporting that they have used AI in the classroom, 
or even in their preparation.
    Access to this technology, and the resources to support 
building AI literacy are unevenly distributed. How we support 
AI in education and its adoption, must be about embracing 
opportunity, access, fairness, especially for under resourced 
rural, urban, and suburban schools and communities. The story 
of AI in education is one of both promise and peril.
    AI offers the promise of personalized learning and improves 
student outcomes. AI's growing capacity to tailor content to 
each student's pace, including those with disabilities, 
represents a significant advancement in education technology. 
However, to unlock personalized learning, AI needs a vast 
amount of data.
    This need can raise significant risks regarding how the 
data is stored, who has access to it, and the potential for its 
misuse. AI promises to enhance the efficiency of educators and 
administrators by supporting them with through a range of time-
consuming administrative tasks. However, tools must be 
reliable, safe, and effective.
    We must work to mitigate the existing biases that exist in 
these tools, increase algorithmic reliability, and trust in the 
outputs. A critical aspect to balancing this promise and peril, 
as my fellow witnesses have said, is AI literacy, and investing 
in professional development for educators, and support for 
students to ethically use this technology.
    America's education is at a critical juncture hindered by 
significant gaps in our own public infrastructure that present 
barriers to innovation that put our economic prosperity, and 
frankly, our national security at risk. Relying solely on State 
level action for AI literacy and AI deployment, is a recipe for 
fragmentation and missed opportunity.
    In education we have seen this before with a patchwork of 
student privacy and data laws that create market failures for 
industry, stifle startups, and limit our ability to harness 
technology to support educators, students, and families. AI's 
nature transcends State lines, making a Federal framework 
critical for interState commerce, and essential for maintaining 
national competitiveness.
    Industry alone cannot carry this burden. Across the board 
cuts to the U.S. Department of Education, National Science 
Foundation, Department of Commerce, and other Federal agencies 
pose a significant threat to our Nation's ability to meet these 
demands, including vital education technology support directly 
to states and districts.
    Imagine an education system empowered by artificial 
intelligence. A force that is transformative as electricity, 
the automobile, the internet. With decisive national action 
America can create robust infrastructure for AI in education, 
unlocking personalized learning, enhancing teacher 
effectiveness, and securing our global leadership by supporting 
schools and districts.
    The stakes could not be higher. Either the United States 
boldly leads, securing prosperity, human flourishing, and 
technological sovereignty for future generations, or we lose a 
race as important to this Nation as the race for space was. 
Thank you.
    [The statement of Ms. Mote follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    

    Chairman Kiley. Thank you very much. Last, I will recognize 
Mr. Chism for your testimony.

  STATEMENT OF MR. CHRIS CHISM, SUPERINTENDENT, PEARL PUBLIC 
              SCHOOL DISTRICT, PEARL, MISSISSIPPI

    Mr. Chism. Chairman Kiley, Ranking Member Bonamici, and the 
rest of the members that are here today, I just want to say 
thank you for the opportunity. It has been a really interesting 
18 hours of getting here. I did not think I was going to make 
it with the flight cancellations, but I am here, and I am 
honored to be here to answer a few questions for you guys.
    I want to take this in a little different direction. You 
have my written testimony, but my written testimony goes all 
over the place. There are 10 or 12 topics there, so I really 
want to cover a couple today that I think are the most 
important. First and foremost, is workforce development. That 
is something that we need to be focusing on in the K-12 world.
    I can tell you this, I did a presentation a couple of weeks 
ago for business and industry. It was open to anybody that 
could come, or who wanted to come. We had people from IT, we 
had people from insurance, we had people from banking, we had 
people from all walks of life coming into this.
    Really, the point of all of this is AI is going to be 
involved in every branch of business moving forward. It does 
not matter where we go or what we do, and that is global. That 
is not just Mississippi, that is not just the U.S., that is 
globally as well. We are going to see AI in everything that we 
do. It is going to be the new industrial revolution.
    A lot of times in education we are the last ones to change. 
You know, it takes us a long time to get things moving, so I 
have really tried to be on the front end of this, in doing a 
lot of training in our State and teaching people how to use 
this, and the right ways to use it.
    The second thing is I really want to touch on efficiency. 
This is a big part of what I do in giving these presentations 
to other school districts, to other states as well. Efficiency 
is where this train is driven from.
    AI can make all of us so much more efficient, and that is 
teachers, that is students, that is administrators, that is 
central office personnel, everything that we do creating our 
own agents that can do things for us is how we create that 
efficiency.
    Again, it is not as hard as people think it is to create 
these things, so in the end, ultimately that's what we are 
trying to do is create efficiency within schools. Again, I 
agree with a lot of things that have been said here already. I 
am not going to rehash a lot of those. A lot of those are in my 
written testimony as well.
    In the end, what we need to do is continue to move forward. 
This is not going anywhere, and so we in the K-12 world, we 
need to respond to this as quickly as possible. I want to end 
with a story quickly, and I have an English teacher who for 29 
years, she can retire anytime that she wants. In fact, she 
called me this fall and said I am going to retire. I said I 
need you to come and talk to me for a few minutes. I do not 
need you to retire.
    She came over and had a discussion with me, and she said I 
just cannot do it anymore. She said I am grading papers every 
night for three and 4 hours a night. It is taking me eight to 
10 days to get this information back to kids. I just cannot do 
it. I need to be a mom. I need to be a wife. I just cannot do 
it anymore.
    I said well, can you spend about an hour with me? She did. 
We created an agent for her. We uploaded the State writing 
rubric, and basically, we took everything that she teaches in 
writing from introductory paragraphs, to conclusions, 
transitions, sentence structure, all of these things, and we 
created an agent for her to use in the classroom.
    Now she does not have to take those papers home and grade 
them, and understand that AI is not doing the grading, that AI 
is a tool that she is using to help grade.
    I think again, that just gives you an example of where we 
can be so much more efficient in the classrooms, as 
administrators, as teachers across the board, and again, I will 
certainly take any questions.
    Again, I am so very proud and honored to be here with you 
guys. Thank you.
    [The statement of Mr. Chism follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    

    Chairman Kiley. Under Committee Rule 9, we will now 
question witnesses under the 5-minute rule, and I will now 
recognize Mr. Rulli for 5 minutes, the Representative from 
Ohio.
    Mr. Rulli. Thank you so much, Chairman. I appreciate that. 
Mr. Chism, I think everyone is a little nervous about AI, just 
in general, but also optimistic for what the future could 
possibly hold. Could you share some of your stories with you 
about students and this toll and ability to excel them as a 
student in America?
    Mr. Chism. Absolutely. I am happy to do that. It really 
begins with the teacher, you know, and again, I do not want you 
to think that I am saying that we need to teach elementary and 
middle school students how to use AI, that is not what we are 
doing. We are really focused on kids that are in 10th, 11th and 
12th grade.
    Again, the idea is to teach them to use this, but also to 
teach them to use this in an ethical way. In our trainings, in 
the things that we do with teachers in training students, what 
we talk about are really two things, and that is if you can use 
this as the perfect assistant, and the perfect search engine.
    You know, you are using it in an ethical way. There is a 
big difference between having an AI program write for you, and 
an AI program write with you. Those are two completely 
different things. What we have seen, especially in the writing. 
I just gave you that example from that teacher, right? You 
know, she can now do these things.
    What I was not paying attention to at the time I was 
helping her, but what we are now doing instead of writing a 
paper every 10 days, we are writing a paper almost every day. 
Again, and then this opens up other things, you know, in some 
of our tech classes as well.
    It gives them new ideas and new ways to take this, and I 
think that is the power of AI, is the idea that it can give you 
ideas that you have not thought up because it has got access to 
the full internet. Again, there are 1,000 stories that I can 
share, but that just gives you an overview of the direction 
that we are going, and the way that we are using it.
    I will tell you it starts with the teachers. It starts with 
administrators and teachers. If they are not onboard first, our 
kids are not going to use it in the right way.
    Mr. Rulli. I really appreciate that. You know, being a 
parent, and also looking back in my own mind's eye, when you 
look at a subject like AI, it is a little contradictory in my 
own mind about the love and the hate and the scariness of it. I 
think, you know, the idea that could cure cancer with AI, you 
could have these kids learn in ways they never possibly could.
    Do you have a little, quick piece of advice for mom and dad 
that maybe are a little bit worried about the guardrails? You 
know, some of us were raised, you know, watching the 
terminator, and I mean that is obviously hopefully not going to 
ever happen, but do you have anything to calm down mom and dad?
    Mr. Chism. You know, I get that a lot in every presentation 
I give is there is a healthy fear, and I think we should have a 
healthy fear of it. You know, I think for parents, it is the 
communication piece with their own kid, and it is making sure 
that they are also teaching good habits to kids with this.
    You know, I think one of the scary things is, and there are 
models out there that are completely open that have no 
guardrails, and I think that is one thing that scares me a 
little bit. You know, you can actually go to Ollama.com and you 
can download any model that you want, and some have the 
guardrails taken off.
    If a kid is in the tech world and knows how to do this, 
that can be a scary proposition for a child. I think again, 
monitoring what your kids do, and having good conversations on 
what good behavior on this looks like is no different than 
social media or using the internet in general.
    Mr. Rulli. I appreciate that also. Finally, probably one of 
the most passionate questions I have for you, I was on the 
school board for 8 years, and I worked with students very 
closely. I was very hands on approach, especially when I was 
the President of the School Board.
    I wanted to take focus on AI, and I want to concentrate on 
an IEP or a 504 Plans, are probably my top priority that I have 
always been in school. I know a lot of moms out there, and dads 
are worried about this. Can you talk about AI, and the 
potential to help our students with disabilities, and how 
exciting that can be for their future in learning?
    Mr. Chism. Oh my goodness, that is fabulous. My wife is 
actually a special education director in the district, so I 
have created several things for our teachers to use, and I 
think the best part of this is you can take a student's data, 
and again this is something we may bring out at some point.
    We have our own AI server in the district, so we do not 
have to worry about FERPA laws, OK. We can take the students' 
assessment scores, and we could match those to State 
assessments, and then we can use the power of these AI's to 
actually script an individual plan for that kid with a path to 
proficiency.
    I think that is the exciting part about this, and literally 
that is done in 20 or 30 seconds. We upload a couple of PDFs; 
we tell it what we want it to do. I have already pretrained 
this in the background to do what I want it to do, and in the 
end, it gives that kid a short-term, medium-term and long-term 
plans that are developed around that individual child, as 
opposed to just the entire classroom.
    Again, that is one of the things that excites me the most 
about this.
    Mr. Rulli. That could probably be even flipped onto like 
higher ed. It could go to the University or whatever, because 
it is so individualized for that particular student. It could 
be a growing tool that that child grows into adulthood with, 
correct?
    Mr. Chism. Yes, sir. 100 percent. I do not think this stops 
at K-12. Honestly, I do not think this stops at college. I 
think we are going to see this moving into the business world. 
Knowledge is going to be something for everybody now, and I 
think that is what is changing in the world of AI.
    You know, we used to pay--we pay doctors lots of money, 
right, but now everybody is going to have access to the 
specialized training that they have had for years, so again, I 
do not think it stops in K12, and college. I think that rolls 
right into adulthood as well.
    Mr. Rulli. How exciting. Thank you. I yield my time back, 
Chair.
    Chairman Kiley. Thank you very much. The Ranking Member, 
Ms. Bonamici of Oregon, is recognized for 5 minutes.
    Ms. Bonamici. Thank you, Mr. Chairman. Ms. Mote, we spoke 
about the Trump administration's plans to dismantle and abolish 
the Department of Education, and what that would mean to not 
only the K-12 system, but institutes of higher education, 
informed policymaking, civil rights, et cetera.
    Could you summarize what, in your opinion, from your work 
as a technologist, what would dismantling the Department of 
Education mean for students, and our locally driven public 
education system, especially regarding the use of AI in K-12 
schools? Then I want to have time for another question too 
about professional development.
    Ms. Mote. Absolutely. Thank you so much. I think we have to 
recognize that as you shared in your opening statements, 
conversations about curriculum, conversations about what is 
happening in the classroom are inherently a local context. The 
Department of Education provides critical support to districts 
across this country, and resources particularly around 
education technology.
    There are over 13,000 public school districts in this 
country. Not all of them can have a Mr. Chism, who builds his 
own AI server. They need cybersecurity expertise. They need 
interoperability and privacy expertise. They need the ability 
to pick up the phone and call someone at the Office of Civil 
Rights and ask questions about whether or not the use of this 
tool is within the bounds of the law.
    We cannot expect that expertise and capacity to be built in 
every school district in this country. If we remove those 
supports, if we remove the critical research and development 
function about what works for whom and under what conditions at 
the U.S. Department of Education, I worry that local schools 
and districts will not actually be empowered to make the 
decisions that they should be making about what is taught, what 
tools, and how we support educators and students throughout 
this country.
    Ms. Bonamici. Thank you. I am sure everyone here recognizes 
the importance of professional development, and you know, 
educators have various levels of skills when it comes to 
technology. I want to ask you. You noted in your testimony that 
at one point there was a risk of overreliance on AI that may in 
fact diminish critical thinking, and creative problem solving.
    I also want to just acknowledge Dr. Dobrin said something 
about the importance of the most sought-after skills, 
communication, critical thinking, problem solving, teamwork. 
How would professional development, how does that make a 
difference, and how would it prevent, or present new 
opportunities for educators?
    How can it inform educators about the risks of AI, and how 
does elimination of the Office of Educational Technology and 
cuts to these professional development programs affect 
opportunities for teachers who are using or want to use more 
AI?
    Ms. Mote. I am going to introduce a critical concept I hope 
to this Committee called, Humans in the Loop, in terms of 
technology. This is the idea that when an IEP is generated by 
AI, there is still a special education teacher, or a special 
education coordinator who is looking at that IEP and saying 
does this match the student that I see in front of me?
    AI is never going to replace the fundamental human 
enterprise that is education, so how do we equip every teacher 
with the necessary literacy skills to work alongside this 
technology, to be stewards of minor data, to protect data 
privacy and security, to choose the right tools, and to stay 
engaged in this technology?
    We can only do that if we support AI literacy and 
professional development for educators throughout this country. 
Let us be honest, when I was running my school, I had a lot on 
my plate. Frankly, like teachers have more on their plate today 
than they did when I was running my school in Brooklyn. There 
are challenges with youth mental health. There are ways to meet 
the workforce demands. There are questions from parents.
    There is increased need to close academic achievement gaps. 
We need to be able to support educators and not just add 
another thing to their plate. They need the learning, they need 
the skills, and they need the understanding of this technology 
if we expect them to adopt it for the efficiencies that we are 
talking about here today, the personalization, but also the 
ability to move forward.
    Ms. Bonamici. Thank you. You mentioned data security and 
privacy, key issues. I also serve on the Committee on Science, 
Space, and Technology, and was a member of the bipartisan AI 
Task Force, and I hope everybody in this building is listening 
to all of the testimony here today.
    Everyone is emphasizing the importance of data security and 
privacy, and I hope that we can get that done on a bipartisan 
basis. I just in my remaining few seconds, I want to ask Dr. 
Rafal-Baer, you talked about the importance of Federal 
research. Where have you seen the most beneficial research 
being done when it comes to artificial intelligence in 
education?
    Ms. Rafal-Baer. I think in the role of Federal research is 
absolutely critical. Federal research helps to signal what is 
important. It provides that scale and ability to look at 
whether things are truly relevant under what conditions they 
are working, and whether or not we are getting the kind of 
return on investment that we would expect from that level of 
research.
    Ms. Bonamici. Where are you seeing that research? The 
beneficial research, what agencies have been doing that?
    Ms. Rafal-Baer. What agencies at the Federal level?
    Ms. Bonamici. Yes.
    Ms. Rafal-Baer. Are doing that? I think a number of 
different agencies on the Federal level who have been doing 
strong research. I think for states, it is about the signaling 
that the Federal Government does in putting out that research 
and being able to stand behind that research.
    Ms. Bonamici. Thank you. As I yield back, Mr. Chairman, 
when it comes to artificial intelligence, I always have 5 hours 
of questions, and I only get 5 minutes, so.
    Chairman Kiley. Thank you. The Representative from 
Pennsylvania, Mr. Thompson, is recognized for 5 minutes.
    Mr. Thompson. Chairman, Ranking Member, thank you so much 
for this hearing. First of all, Ms. Mote, I want to say I 
couldn't agree with you more. I mean AI is an incredible 
opportunity, but it does not replace our teachers. It makes 
them more effective.
    AI without either HI or RI, human intelligence, or real 
intelligence, is just not going to get it done, and it may 
mislead us. As we have heard today, artificial intelligence 
presents a unique opportunity to rapidly advance the 
educational landscape for students with disabilities. In just 
the last two decades we have seen enormous growth and success 
of assistive technology in aiding these students, and I believe 
AI presents the next frontier in helping to ensure all students 
have access to a high-quality, appropriate education that bests 
fits their abilities.
    Maybe that is why we have not gotten around to doing our 
job of reauthorizing Individuals with Disability Education Act. 
You know because with this new frontier, we would be taking 
into consideration, as we actually do the reauthorization of 
IDEA, it is long overdue.
    Dr. Rafal-Baer, I know AI is making possible the 
development of new screening tools and assistive technologies, 
such as screen readers. Given your previous experience as a 
special education teacher, and thank you for doing that, could 
you talk more about the benefits AI adoption might have for 
students with special needs?
    Ms. Rafal-Baer. Yes. Thank you for that question. I am 
really excited as a former special education teacher about some 
of the power and the potential, particularly when it comes to 
things like screeners. I think there are so many opportunities 
to learn more about our students, and to be able to use that 
information in connected ways to give us a fuller picture about 
each individual student, helping to unlock new ways to think 
about supporting them.
    I also think that like all things, this is very new, and I 
think being able to really understand and look at the impact of 
this, and see whether or not these screeners, and these tools 
are working the way that we would expect, and for what types of 
students, and under what conditions is such an important 
research function that we need to continue to explore.
    I think that there has been some really promising early 
research recently out of Ector County in Texas, where they were 
looking at the way in which AI could support not the students, 
but the actual tutors, the individuals who are going in and 
providing support to those students, and looking at tutors at 
different skills level.
    Those who were lower on their skill level, those who were 
medium in their skill level, and those that were said to be 
higher performing. What they found was that AI supported all of 
them, but particularly those that were lower and middle 
performing.
    As a special education teacher, that is really exciting. 
That means that once we have that information from these 
screeners, and we can better pinpoint areas that we need to 
support our students, that there are promising and encouraging 
results from the research around how AI might help all of the 
adults who were working with our students, but particularly 
those that need more support themselves to better differentiate 
and scaffold for those students, and that is something that 
really excites me.
    Mr. Thompson. Do you believe that both teachers and schools 
at this point have the resources they need to ensure students 
with disabilities can take advantage of these potential 
benefits?
    Ms. Rafal-Baer. I think the resource question is always a 
really tricky one. There is a role at every level in terms of 
resources. There is a Federal role. There are states 
prioritizing resources. I think the private sector has a really 
critical function to play in this world of AI. I think being 
able to bring everyone together, and doing so in ways where 
they are partnering effectively, and sharing research and 
information is a vital function, and it is a vital function of 
our states.
    It is imperative that our states are leading on this, and 
that our states are collecting data and information, and 
reporting transparently and publicly so that individuals 
understand what this looks like, and how those resources are 
being used and to what impact.
    Mr. Thompson. Well, with benefit usually comes some type of 
risk, and so for example in managing the risk of AI generated 
images, the spread of bullying tactics, or even AI that has not 
been properly programed to work with specific learning 
disabilities, could impact special needs students more than 
others.
    I look forward to talking with you offline because I am 
going to run out of time here but basically looking to see what 
risk do you see AI adoption posing, particularly for those 
students who you would advise school leaders to balance those 
risks with the AI benefits. I think that is a conversation that 
we need to have as we pursue this, so.
    Ms. Rafal-Baer. I look forward to speaking with you further 
about this. I think it is one of the most important and 
critical topics that you all could undertake.
    Mr. Thompson. Thank you so much, and Chairman, I yield 
back.
    Chairman Kiley. The Representative from New York, Mr. 
Mannion, is recognized for 5 minutes.
    Mr. Mannion. Thank you, Chairman Kiley, and Ranking Member 
Bonamici, for holding this hearing today. Thank you to the 
witnesses for your detailed testimony and recommendations. AI 
holds potential in the classroom and outside of the classroom, 
and it can improve, I believe, you know, K12 education, and be 
an important tool for teachers that utilize it, particularly as 
it relates to students entering the workforce.
    I, of course, share the same concerns of many members of 
this Committee, that we must be thoughtful in our approach, and 
that the opportunities that come with AI, also come with a 
variety of risks. As we further integrate AI into the 
classroom, we have to ensure that it supports and not supplants 
the essential roles of teachers, and the broader purpose of 
education.
    As a former teacher, I believe that it is much more than 
academics. It is about developing well rounded individuals that 
can think critically, constructively engage with their peers, 
and communities, and successfully navigate an ever-complex 
world. My question is for Ms. Mote.
    Your testimony mentions a common concern about AI, that if 
not used properly may negatively impact students' social, 
emotional development. Could you elaborate on that please?
    Ms. Mote. Absolutely. I think all the witnesses have talked 
about the skills that employers are going to demand in an AI 
empowered world, critical thinking, working in multi-functional 
teams, thinking about how to do creative problem solving.
    That happens when you work alongside, and with humans. I am 
just going to ask you about the future that we might want. 
Recently a Stanford study found that a tool called Replica AI, 
which allows young people to create a digital twin to engage in 
conversations with reduced suicidal ideation by 3 percent. That 
is great, 3 percent.
    If we get any change in young people's mental health and 
suicidal ideations that is great. Last year the CDC released a 
study which said one meaningful connection with a human in a 
school, a bus driver, a teacher, a coach, reduces suicidal 
ideation in young people by 33 percent.
    Do we want 3 percent or 33 percent? I would prefer for my 
young children, to have those meaningful connections with 
adults who are helping them navigate this technology, work 
alongside it, and thinking about how to have those skills that 
employers will demand.
    Mr. Mannion. Thank you for that. My followup question was 
actually going to be about the identification of suicidal 
warning signs, so I think I am going to leave your testimony 
right there and yield back to the Chair.
    Chairman Kiley. The Representative from Indiana, Mr. 
Messmer is recognized for 5 minutes.
    Mr. Messmer. Thank you, Mr. Chairman. Dr. Rafal-Baer, in 
your written testimony you said that states are ideally 
positioned to support districts in making locally grounded 
decisions that align with their unique goals. You touched on 
this during your testimony, but what role do you think states 
can play uniquely in an AI adoption where individual districts 
might struggle?
    Ms. Rafal-Baer. Thank you for that question. States are 
first and foremost, the ones who set the vision for where 
things are going within their State, and they do so through 
deep community engagement, and understanding where their 
constituents are. That is only more true in this moment around 
AI.
    I think states also play a critical data collection role. 
We saw this with internet access, and the differences in the 
kind of quality of access and the unevenness between what 
students had access to, even within a school setting versus 
what they might have at home or an after school.
    We need to look at that in the same way with AI. We know 
that students are going to have very different experiences, and 
we are already seeing that in early survey work where we are 
seeing higher socioeconomic students, who have more access to 
paid versions of AI. Those are important things that states 
should be looking into and should be collecting data about.
    The second area that we suggest in ILO Group's framework 
for State education agencies is the concept of an AI assurance 
lab. We think states are uniquely positioned to set up a lab 
that creates public transparency around what tools are being 
vetted by scientists and researchers from within the State, 
thinking about bringing task forces with other representatives 
on it, including having parents and caregivers a part of those 
labs.
    Not just doing the approvals on the front end, but also 
rigorously and often doing more research on the outputs of 
tools that are approved can help to give people more confidence 
and just understanding about all of this, and we think it is an 
important role that states can play.
    Mr. Messmer. Thank you. Dr. Dobrin, you mentioned that 
about 25 states have, or are developing, official guidelines 
and policies for K-12 AI education. On a whole, do you think 
these states are rising to the challenge, and giving the 
districts the information they need, or do these guidelines 
lack specificity and clarity?
    Mr. Dobrin. Thank you for that question. In regarding what 
different state's legislation, like I had mentioned in the 
written, there are about 45 states now that are considering AI. 
Most of those are not directed specifically at education. 
However, the trickle down will inevitably be that that will 
affect education. The thing that I think that comes out clear, 
and I can point to yesterday's Florida Senate discussion in the 
Commerce Committee about AI, is that the majority of these 
states are focusing on further research, which is, as my 
panelists have--other witnesses have noted, that is the 
important part at this point.
    We need more data before we make these decisions. Those 
states that are beginning to really look at what AI deployment 
looks like, both in workplace, and in education, they are the 
ones who are going to--excuse me, they are the ones who are 
going to lead that conversation.
    Mr. Messmer. Have there been pretty robust collaborative 
efforts between states that you've seen?
    Mr. Dobrin. I honestly do not know how states are talking 
to other states.
    Mr. Messmer. Thank you. I yield back the rest of my time.
    Chairman Kiley. The Representative from Connecticut, 
Representative Hayes is recognized for 5 minutes.
    Mrs. Hayes. Thank you. Thank you to our witnesses for being 
here today to testify. Advanced instructional technologies can 
be transformative, enhancing both teaching and learning. 
However, any implementation of AI technology in the classroom 
should maximize both safety and privacy, while also advancing 
equity and fairness.
    Ms. Mote, recently the Office of Education Technology, or 
OET, at the Department of Education was eliminated as part of a 
reduction in force at the agency. OET played an integral role 
in developing and implementing policies and initiatives on 
education technology. What has been the role of OET, and how 
has this office assisted states in the implementation of AI 
technologies in the classroom from your perspective?
    Ms. Mote. Well, I think it is important to name that OET 
really puts out guidelines, and really best practices for 
states and districts to follow, providing that necessary 
expertise and knowledge that frankly would be impossible for 
there to be an AI expert in every district to be able to do.
    Ways to communicate to parents who say that their schools 
right now are not communicating with them about AI. 83 percent 
of parents have said in a commonsense survey that they have not 
heard anything from their schools about generative AI. As a 
parent, we know that we need to be working alongside our 
educators to provide that critical guidance.
    I want to talk to you about something very quickly, which 
is a foundational layer of data privacy, security and inter-
operability, which every technology needs, whether it is AI, or 
it is cybersecurity. From 2002, sorry from 2022 to 2023, there 
was a 105 percent increase in known ransomware attacks against 
K-12 and higher education institutions.
    State and local cybersecurity grant programs were provided 
not only by the Department of Education, but also by CISA and 
other entities that provided critical support for not just AI, 
but technology as a whole. Without that guidance, without that 
knowledge, without that expertise, states and districts will 
have to navigate this changing technology on their own.
    Mrs. Hayes. Thank you. I appreciate that. To your point it 
is very important. This is something I have been saying all 
month that people understand the role of the Department of 
Education in providing frameworks and guidelines. These things 
are implemented at the local level, but in many districts, they 
do not have the resources or the knowledge, or even basically 
the skill set to pursue aggressively emerging technologies and 
things like that, especially in a field that is moving as 
quickly as AI.
    Earlier this month the Department of Education eliminated 
half of their staff. Additionally, President Trump signed an 
executive order calling for the dismantling of the Department. 
Ms. Mote, again, how does the elimination of OET, and the cost 
to dismantle the Department entirely, impact the role of the 
Federal Government in supporting access and deployment of safe 
AI technologies in the classroom?
    Ms. Mote. Well, I think everyone here has talked about 
access, and how important it is for us to be thinking about 
access. I want to think about the role of the essential 
frameworks and connectivity that have powered our communities, 
and the ability to use that technology.
    The U.S. Department of Education has provided that support. 
The U.S. Department of Commerce has provided that support. CISA 
has provided that support. We talk about the need for states to 
be able to drive connectivity, to drive access, but they would 
not have been able to do that without universal service funds.
    They would not have been able to do that without the 
Digital Equity Act, Title II, Title IV dollars, in order to 
lead professional development. If we really want there to be 
the type of infrastructure that allows all students to thrive, 
we need to invest in institutions that can provide that 
capacity.
    Mrs. Hayes. Thank you. I appreciate that because I can tell 
you even within the school district where I worked in my 
school, our computer and technology classes were taught by 
teachers, not practicing professionals in the field, so the 
ability to have outsiders come in with additional support, or 
professional development, which is something that the 
Department of Education provides for school districts across 
the country is not only critical and fundamental, but it is 
necessary.
    Especially when we are talking about something like AI 
where the technology is advancing so quickly. The safety of our 
students has to be a critical focal point. I do not have much 
time, but I am going to ask a question. If maybe you could 
submit the answer for the record, that would be great. Just 
some ideas on steps that local school districts can take when 
vetting AI technology to ensure that student access is tailored 
to educational content, and improves engagement, but also does 
not allow disruptions in learning, or many of the other things 
that we saw.
    Thank you, and with that I yield back.
    Chairman Kiley. The Representative from Utah, Mr. Owens, is 
recognized for 5 minutes.
    Mr. Owens. Thank you. I want to first of all thank the 
witnesses for participating in this very timely use of AI in 
education. If we want our children and our education to be the 
forefront of the world, we have to start adapting the latest 
technologies of use in classrooms in a safe, responsible 
manner, and I look forward to again, talking with the witnesses 
about this.
    It is probably one of the more important conversations we 
are going to have in a bipartisan way. Dr. Dobrin, one trick 
aspect of AI adoption in schools is assessing the cognitive 
impact. You mentioned that further research might be helpful to 
understand how AI generated content and automated problem 
solving might affect a student's ability to think critically.
    That has a real concern. What is obviously a rapidly 
growing field, how does research based tell us so far of AI's 
impact on critical thinking, and what type of additional 
research would be helpful?
    Mr. Dobrin. Thank you for that. Yes, indeed, that as other 
witnesses have said, research right now is one of the keys. 
What we do not have regarding cognitive development right now 
is long-term affect research. We have data that both shows the 
improvement in critical thinking and problem-solving skills 
when AI is in use, and we also have data that also shows that 
some students are turning to the technology to do the work for 
them.
    I think that the real key here for educators has a lot to 
do with demystifying how these technologies actually work. In 
the conversation we have had thus far, we tend to talk about AI 
as a single thing. AI does this, AI can do this. AI is an 
umbrella term for a lot of different kinds of technologies.
    When we start looking specifically at cognitive 
development, part of what we have to think about is how we are 
using those various kinds of AIs in various context. This is 
one of the reasons that in my work, and this was also brought 
up by one of the other witnesses earlier, that I emphasize that 
this is about augmentation, not automation.
    All AI--all AI requires three things, fantastic computing 
process, lots of data and human expertise, what was referred to 
as human in the loop a moment ago. If we can influence that, 
then part of what we are inherently influencing also is that 
role of cognitive development. That the human is the one 
running the process.
    That to me is where the research has to unfold, is how that 
expertise gets manifest.
    Mr. Owens. OK. Thank you so much. Dr. Rafal-Baer, you 
highlighted the importance of State autonomy and implanting AI 
literacy and integrating AI tools into the K through 12 
education. As AI adoption grows across school systems, how can 
we ensure that all states can develop the necessary frameworks 
and support structures to ensure the safe and effective AI 
integration?
    Ms. Rafal-Baer. Well, ILO Group has built though two 
frameworks. One that has been developed for State education 
agencies, and one for district superintendents. That was based 
off work that we have been doing with the working group of 40 
different districts and states across the country.
    Beyond creating these frameworks, which are grounded within 
the political operational, technical and fiscal realities of AI 
implementation, and giving department by department questions 
to start to consider, we have also rolled out a three-tiered 
process that begins with deep stakeholder engagement.
    Our belief is that there is no work that should be getting 
done on AI that is not starting first with an incredible amount 
of listening within your community and within your State. We 
think that you go from that part of the process to developing 
ethical guardrails and principles based on the kinds of 
conversations that are coming up in your community, using that 
to then develop guidance.
    Then from there you can go into larger scale optimization. 
I think what ends up happening so often though is that states 
start to feel pressure and in turn, districts start to feel 
pressure that they need to do something. They need to show that 
they are being responsive.
    Well intentioned organizations will provide templates of 
resources, and then often you will hear all these data points 
that, you know, 90 percent of states have guidance that is up 
on their websites, but it has not trickled down.
    To the comment that was made earlier about the reality that 
many parents do not feel like they are even hearing about AI, 
that is because there are too many templatized guidance and 
resources being provided without starting with deep stakeholder 
engagement, and listening, and bringing these conversations to 
our communities.
    We are hearing over and over again from the leading 
providers that we are moving closer to AGI, right? The idea 
that artificial intelligence will have human-like capabilities. 
That means that we have to be doing even more around 
stakeholder engagement and communication to make sure that all 
parents and citizens really understand that.
    Mr. Owens. Well, thank you. Thanks so much. I just wanted, 
and I am running out of time here, I just want to say real 
quick, to the point of the Department of Education. That ship 
has sailed. We now make sure that innovators like yourself are 
part of this conversation, and not bureaucrats of D.C. I am 
excited about the fact we are finally going to be putting this 
power back in the hands of parents and schoolteachers, and 
those in local levels, so I am excited about that process. With 
that, I yield back.
    Chairman Kiley. Representative Lee of Pennsylvania is 
recognized for 5 minutes.
    Ms. Lee. Thank you, Mr. Chairman. Thank you so much to the 
panel for your testimony. Artificial intelligence and its 
algorithms are built by humans, and they are trained on data 
that is created by humans, which means it should not come as a 
surprise to anyone that AI often reinforces the same implicit 
and explicit biases that humans carry, classism, or sexism, 
racism, ableism.
    I just want to really lift up, that we cannot ignore the 
capacity of AI to deepen the systemic inequalities already 
facing marginalized communities, and AI in K through 12 schools 
is not any different. With a dismantled Department of 
Education, where a drastically reduced Office of Civil Rights, 
or Office for Civil Rights is being weaponized to undercut 
students' civil rights.
    K to 12 schools are unprepared to address the vast 
inequities that can come with AI. We should be especially 
concerned about rapidly expanding school surveillance AI, that 
includes facial recognition technology with error rates of 35 
percent for black girls, an online chat monitoring and so-
called aggression detection that is layered with racialized 
assumptions about communication.
    Much of this surveillance technology is an extension of 
histories of scientific racism and eugenics. Ms. Mote, in your 
testimony you referenced a resource created last year by the 
Department of Education on avoiding discriminatory use of 
artificial intelligence.
    The Department of Education, under the last administration, 
created this resource to inform schools of how AI can violate 
students' civil rights. It included examples like facial 
recognition, consistently misidentifying black students and 
predicted algorithms flagging black students as more likely to 
carry out disciplinary infractions.
    Without resources like this from the Department of 
Education, are schools equipped to keep students' civil rights 
intact when deploying AI surveillance?
    Ms. Mote. It is important to name that there is no tool as 
you mentioned in education that is without bias. Every tool in 
use in America's classrooms today has algorithmic in 
reliability or bias built in because it is built by humans. As 
you so aptly mentioned, it repeats the bias that is already in 
existing datasets, or the bias that is held by the developers 
themselves.
    There are tools to mitigate this effect, something called 
reweighting, which allows you to actually index for students 
with disabilities, or students from subgroups. In order to do 
that, to have the data to be able to train those models to be 
more equitable, to be more fair, to offer opportunity for every 
student, you need data and data infrastructure.
    Right now, we are seeing a dismantling of our data 
infrastructure at the Federal level. The very datasets that 
would allow industry, that would allow researchers, that would 
allow others to use that data to be able to train these tools 
to mitigate bias. It is really important that we name that 
there is no way to eliminate bias, but we can provide guidance, 
like the guidance you mentioned, provide data for training, and 
work alongside industry to mitigate bias that exists in these 
tools. Yes.
    Ms. Lee. Thank you. The fact that we can no longer rely on 
the Department of Education to serve its key functions, 
including its jobs of safeguarding students against 
discriminatory AI surveillance is especially concerning, since 
school surveillance technology is a--multibillion-dollar 
industry that seems to really allude oversight and 
accountability.
    Ms. Mote, do the companies selling their AI surveillance to 
schools typically evaluate their algorithms for racial bias? 
How important is it that these companies evaluate and monitor 
products to mitigate that bias?
    Ms. Mote. One challenge for all companies that are using AI 
right now is they can only look at the data that they have 
themselves. Let me go back to my days of making copies in the 
copy room. Has anyone ever made a xerox of paper, and then you 
make a xerox of it again, and then you make a xerox of it 
again?
    Over time that copy degrades. If you only test your tool 
against the data that you have yourself, it is not going to be 
able to surface the bias, or to surface the challenges that 
might be within that tool. We need to provide an infrastructure 
for industry to walk alongside schools and districts, to be 
able to make sure their tools are equitable.
    Ms. Lee. Thank you. I just want to really quickly, in the 
2023 survey by the Center for Democracy and Technology, almost 
40 percent of teachers reported that sensitive student data was 
being shared with law enforcement and being used in predictive 
policing algorithms who identified children who may commit 
future criminal behavior.
    This was closer to 50 percent of teachers at schools 
receiving Title I funding, and closer to 60 percent of special 
education teachers. I wanted to mention that before I have to 
just conclude because without guidance, accountability, and 
oversight from a functioning Department of Education, AI 
surveillance in schools will exacerbate the school to prison 
pipeline and cause irreversible harm to black, brown, and other 
vulnerable communities.
    It is why I introduced a piece of legislation last year 
that will require the Federal agency using AI to establish a 
dedicated civil rights office to identify, prevent, and address 
algorithmic biases, but that is just scratching the surface. We 
know that there is so much more to do, so I thank you so much 
for your time, and I yield back.
    Chairman Kiley. I will now recognize myself for 5 minutes. 
One of the reasons that I wanted to hold this hearing on AI in 
education is not only because I think the potential is so 
incredible in a specific context of education, but also because 
I think it offers a window into just how transformational the 
technological changes that we are experiencing right now are 
going to be across all sectors of society, which I think is 
something that really policymakers and the public at large is 
starting to get--to wake up to a little bit.
    That there has been, you know, some references to that 
during today's hearing, specifically Ms. Mote, you mentioned 
this is an technology on par with electricity. Ms. Rafal-Baer, 
am I saying that correctly? Close enough, Rafal-Baer, you 
mentioned that the leading labs are increasingly talking about 
the arrival of AGI, artificial general intelligence.
    I wanted to give each of you a few moments to expound on 
those ideas.
    Ms. Mote. Well, an arrival technology, right? This is 
electricity. This is the internet. This is automobiles. You 
used to have to pick up the phone and call a dispatcher to call 
a taxi. Now you can do it from your phone. Arrival technologies 
displace workers. They can require new skills and competencies, 
but they also require a level of support in disrupting of 
existing systems, so that they can be embraced across industry, 
rather than having them replace humans.
    I think the work that is happening right now in terms of 
what Congress can do, our colleagues have really mentioned the 
need for research and development, the support for public data 
and infrastructure, the need for some set of guidelines and 
guardrails, supporting educator professional development, 
providing equitable access.
    Mr. Kiley. Right.
    Ms. Mote. This body really convening collaboration and 
dialog. If we are to meet this moment, we need to do it 
together. I look forward to the opportunity.
    Mr. Kiley. Great, thank you. You bet.
    Ms. Rafal-Baer. Thank you for the opportunity to share more 
on this. I think that the fact that we are hearing over and 
over again that we are getting closer and closer to AGI is 
something that concerns me a great deal that this is not a 
conversation that is more commonplace across America right now.
    The fact that once again we are in a moment where it feels 
like we are asking our education system to take on something 
that is actually a societal issue, and I see this, and hear 
this every single day from State and district superintendents 
who earnestly want to move forward on this work, but are really 
concerned about cybersecurity and data.
    They need to have far more expertise, and that is a 
critical role of the Federal Government. That is a critical 
role of a role like a national cybersecurity director, who 
should be helping to make sure that we have the safest systems 
possible around cyber and data, so that states and districts 
can do their correct role, which is to lead these conversations 
locally.
    To make sure that they have visions around artificial 
intelligence, and the kind of workforce upskilling that is 
required that matches their local context and their local needs 
with them being able to take off their plate this level of 
expertise and oversight around keeping their systems cyber free 
in terms of cybersecurity issues.
    Chairman Kiley. Thanks very much. Mr. Chism and Dr. Dobrin, 
I wanted, if we have time permitting, to pose a question to you 
as well. There has been a lot of discussion about further 
research that is needed, some of which has already happened. If 
you could pose one kind of urgent research question when it 
comes to evaluating the effectiveness of certain AI tools, what 
would it be?
    Like one possibility for me would be the immersiveness of 
the tool. You know, you could have a sort of animated Einstein 
teach you physics. You could have a text-based Einstein teach 
you physics. If we have sufficiently developed VR, you could 
actually float through the cosmos as it happens, so that might 
be one area to look at.
    What would be the research questions that are most urgent?
    Mr. Chism. I will go ahead. I really think I love the idea, 
especially with students, but I am really focused on the 
teacher side of things because we have to have teachers to get 
to students. For me it is really teacher efficacy. I think that 
is the big thing. How can we use this to make them better and 
more efficient at what they do?
    Again, I do not have that 28, 29, 30-year veteran that 
wants to walk out the door. She wants to stay, you know, the 
teacher that I have talked about earlier, she is staying with 
us. She is not going to retire. To me it is about teacher 
efficacy, and I think that is the primary focus for me.
    Chairman Kiley. Great.
    Dr. Dobrin.
    Mr. Dobrin. I would pick up on something we have already 
been discussing, and that is access. I think there needs to be 
longer term research on not just where access is available. For 
instance, in Appalachia in the Southwest, where broadband is 
not as easily accessible, but more so what happens in the 
distinction between our students, who only have access to the 
free datasets, the free LLMs, GPT, Gemini CoPilot as opposed to 
those students who have the socioeconomic access to proprietary 
datasets, and to platforms that are more disciplinarily 
specific.
    We need to figure out the differences of what happens with 
those access points.
    Chairman Kiley. That is an interesting question. I mean one 
of the things we see happening though is that what is the paid 
tier this month is available to the free tier next month, so 
there is this sort of, you know, democratizing aspect to the 
technology. With that, the Representative from Florida, Ms. 
Wilson, is recognized for 5 minutes.
    Ms. Wilson. Thank you, Chairman Kiley, and Ranking Member 
Bonamici, for today's hearing. Thank you so much, and thank you 
to our witnesses for being here today. As you know, public 
schools are the backbone of our democracy. As AI continues to 
be implemented in our education system, we must work in 
addressing its potential and worry about the disparities that 
come with this implementation.
    Clearly, education has some big challenges with AI, but 
this Committee is missing the real crisis. The dismantling of 
the Department of Education. It is absurd to envision a bright 
future for our students when the Office of Education 
Technology, vital for AI oversight, has just been shut down.
    It is even more absurd for the Subcommittee to ignore the 
wrecking ball crushing into our public-school systems across 
the Nation right now by abolishing the Department of Education. 
This is like worrying about the ship's Wi-Fi access while the 
Titanic is sinking.
    With this, I have a few questions. Ms. Mote, can AI work 
without human input? What are the consequences of gutting the 
very workforce responsible for ensuring AI is used ethically 
and effectively?
    Ms. Mote. Can AI work without human input? No. In fact, 
generative AI and artificial intelligence systems have to be 
trained on both our responses, and the data that is being used. 
Now, let us just talk about what happens when we take away data 
that could be used to train these systems effectively alongside 
humans?
    Right now, we are seeing families lose access to 
information on the college and career scorecard, being able to 
make critical decisions about what is happening, or what 
choices they make with the loss of IPED's data, or the loss of 
Ed Facts data, where if you are moving from one community to 
the other, not being able to know the quality of your schools, 
or how to compare one school to the other.
    This vital support that is provided, not just by the humans 
who are helping guide local states, excuse me, local districts 
and states, but also that are guiding families through the 
provision of open access datasets, and knowledge is a critical 
role that if we lose our families will be less informed, our 
students will be less prepared, and our educators will be left 
without the resources they need.
    Ms. Wilson. Ms. Mote, with the Office of Civil Rights being 
sidelined or weakened, how are students, particularly those 
from marginalized communities supposed to ensure their rights 
are protected if AI tools amplify existing disparities because 
they will?
    Ms. Mote. I want to return to--I do not think an educator 
wants to harm a student, but they need to know about these 
tools. They need to know about the tools that are being used. 
They need to have research about what is the most effective 
tool in order to make the right choice for a student.
    Every educator in this country wants students to succeed. I 
know that because I meet them every single day, but they do not 
necessarily have the expertise to know that this tool or that 
tool is better, that the $99.00 edition of Khanmigo is better 
than the free tutor, and they do not have the resources to make 
those choices, and to navigate those deep, deep challenges that 
they have to meet young people.
    If we want to overcome achievement gaps, we need to make 
sure that our educators have the knowledge, expertise and 
capacity to make the right choices for students alongside 
families. That can only happen when we have that knowledge, 
when we have that capacity, that data, and that expertise.
    Ms. Wilson. Ms. Mote, how do we prevent AI from becoming 
just another tool that benefits already advantaged, and leaves 
behind the same black and brown students who have historically 
been underserved and who have no computers?
    Ms. Mote. Thank you. I think equitable access is something 
every single witness has talked about, and so I think Congress 
needs to renew their investment in public connectivity 
infrastructure. They need to have targeted programs to ensure 
that all communities have access to the technology, the 
hardware, reliable internet connectivity, and the adequate 
technical support to benefit every community in this country.
    Ms. Wilson. Thank you. I yield back.
    Chairman Kiley. The Representative from North Carolina, Mr. 
Harris, is recognized for 5 minutes.
    Mr. Harris. Thank you, Mr., Chairman, and thank you to all 
of you on the panel for the incredible testimony you have given 
today. I appreciated the opportunity to read through your 
testimony that was submitted to us prior. I have a couple of 
questions, and really the first one is going to be to you, Mr. 
Chism.
    Your written testimony mentioned that an accurate or biased 
content is a big challenge with AI adoption. As you pointed 
out, AI models are only as good as the underlying data that 
they are trained on, and that is what concerns me.
    Research from Carnegie Mellon University revealed that 
large language model bots trained on the internet since 2016, 
show more polarization than bots trained before Donald Trump's 
first election. In fact, the research also showed that bots 
trained from books were more socially conservative than bots 
trained through the internet or social media.
    There is in fact numerous examples of left leaning 
political bias that shows up when asked if a white Christian 
man should be ashamed, Google Gemini lists a variety of liberal 
buzz words like, ``systemic injustices,'' and ``marginalized 
communities.'' When asked if a black female lesbian should be 
ashamed it says, ``Absolutely not.''
    The free version of ChatGPT is unable to acknowledge Donald 
Trump as the current President. It even says that Joe Biden is 
in the White House in 2025. Now, I point all that out, Mr. 
Chism, to come back to you as superintendent of a school 
district there in Pearl, Mississippi, how do you make sure in 
that role that your district's AI use does not amplify existing 
biases, or spread false information?
    Mr. Chism. I think that is a fabulous question, and that 
has got a lot of answers to it. I will tell you that some of 
the big companies are working on this. I mean they do realize 
that some of that bias is built in. In the end, I mean there is 
going to be bias in every computer program that you create. For 
us as a district we have created our own server, so we get to 
do the training ourselves.
    I think that is one big advantage for us, and I will say we 
discussed equity as well. We are actually working with a huge 
internet company in Mississippi, Ceasefire, and we are actually 
working on trying to make our model available for all school 
districts in the State of Mississippi completely free. We are 
working on that.
    Again, is there a way to script it all out? Absolutely not. 
Not yet, but I think we are moving forward to that, so I think 
it is good training, and it is really having--its good 
communication. It is people, when they see those things happen, 
it is pointing those things out, and then we can go back and 
make adjustments to that model, even on our end, on our server.
    I can tell you again, that even using the model such as 
ChatGPT, they know that this is a problem, and they are working 
on it. I am on a computer call with them about every 2 weeks 
with Open AI, so this has been a discussion that we have had. I 
do not think there is a perfect answer for it, but I do know 
that they are working on it, and certainly that is something 
that we will monitor 24 hours a day because we want to make 
sure that the information that we get back is good.
    I will also say, even using the larger language models, 
like ChatGPT if you are on the paid version, or even the free 
version of that. You can go in and script some of this out 
yourself. You know, if you tell it how you want it to respond, 
you tell it the websites that you would like for it to go to, 
to look for information.
    You know, I have that scripted into mine, so now I know 
where that information is coming from, so I realized that now 
that information that is coming back is not from Wikipedia. It 
is good information from the sites that I determined that I 
want it to come from.
    There are ways around that, even though the model itself 
does not fix it.
    Mr. Harris. Are you finding that schools should be weary of 
the development behind the AI bots, and be selective about the 
programs they use in their schools?
    Mr. Chism. Oh, absolutely. 100 percent. I think, going 
through all of this information, and getting really good 
information. Again, if we were to go with a company, a large 
company, that would be a multi-week investigation into what 
they do and how they do it before I put it in front of kids.
    Mr. Harris. Got you. Thank you, sir. Dr. Rafal-Baer, in 
your recommendations for policymakers, you mention that the 
Federal role should be intentionally limited. I agree. 
Especially since this administration, and obviously the 
direction we are moving will be working to eliminate the 
Federal role in education entirely.
    What would you say to someone who argues that the Federal 
role should be unlimited, rather than limited, and what kind of 
detrimental effects would overregulation have on schools?
    Ms. Rafal-Baer. Well, I think we have been talking about 
this so much in this Committee. This is moving too fast. There 
is nothing that the Federal Government would put on paper that 
is not going to be outdated, even a couple of months from now. 
It is imperative that states have the ability and the 
flexibility to make these decisions within their own context 
and allowing districts the ability to innovate.
    It is critical that the Federal Government not play a role 
in defining things like AI literacy, or AI curriculum. 
Curriculum and content all should be left to local levels. 
However, I think the place that is really critical is around 
the cybersecurity and data privacy.
    Mr. Harris. OK. Well, thank you. Mr. Chairman, with that I 
yield back.
    Chairman Kiley. The Representative from North Carolina, Ms. 
Adams, is recognized for 5 minutes.
    Ms. Adams. Thank you, Mr. Chairman, and thank you to our 
witnesses for being here today. Last month, Charlette-
Mecklenburg Schools, one of the largest districts in North 
Carolina, launched a bold, district wide effort to shape a 
community vision for AI in education.
    They are inviting parents and teachers and students to the 
table, not just to respond to change, but to lead it. That is 
the kind of leadership that we need. Instead, the leadership 
that we are seeing at the Federal level is tying our schools' 
hands with budget cuts, with confusion, with political agendas 
that have absolutely nothing to do with what is best for our 
most disadvantaged kids.
    The kids on free and reduced lunch, the kids who rely on 
Medicaid, the kids whose only option is public school. AI is 
already reshaping our classrooms. The real question is how are 
we helping schools use it responsibly, or are we leaving them 
to figure it out on their own?
    Here is what I have been hearing from educators, from 
parents, and my own daughter, who was a public-school 
principal. The promise of AI is real, but so are the risks. We 
have already seen cases, whereas I reiterate my colleague 
Representative Lee said, AI tools flag black parents for 
plagiarism at higher rates than their peers.
    That is not innovation, that is bias, plain and simple. Ms. 
Mote, AI is not just showing up in lesson plans. It is now 
embedded in how students are evaluated, how feedback is given, 
and how learning is tracked. How can schools ensure that AI 
tools are actually supporting equity. I know some people think 
that is a bad word, I think it is a good one, in classrooms not 
reinforcing bias?
    What role should the Federal Government play in setting 
guardrails to make sure that these schools serve all students 
fairly?
    Ms. Mote.
    Ms. Mote. Thank you. You actually offered the answer when 
you were talking about the great work that Charlotte-
Mecklenburg is doing right now, which is building AI literacy 
with parents, with communities, with educators, and really 
putting students and parents and communities and educators at 
the center of the work that they are doing.
    You have to be able to ask questions of these tools. You 
have to be able to understand if the results that they are 
putting out from these models have inaccurate information or 
might not potentially have the right types of inputs. If we do 
not build AI literacy in our educators with our parents and 
with our communities, we are not going to be able to use these 
tools to the best of their advantage.
    Ms. Adams. Well, thank you. Thank you for your great 
comments about Charlotte Public Schools.
    Ms. Mote. They are wonderful down there at Charlotte-
Mecklenburg.
    Ms. Adams. Our Chair of our School Board is sitting out 
here, and I am happy to see her. While Charlotte public schools 
are trying to move forward with innovation, we are facing real 
setbacks. Just last month we lost over 5 million dollars in 
Federal grants.
    Funding that helped high need schools recruit and train 
diverse teachers, and those grants were simply cut because they 
supported DEI. I still think those are good words. Let us be 
clear. Students learn best from teachers who understand them, 
who reflect their community.
    It is that kind of representation that is not a luxury in a 
district where students represent more than 175 countries, it 
is essential. Ms. Mote, we know AI is a powerful tool, but it 
does not replace teachers, especially not teachers who reflect 
the lives and identities of the students they serve.
    What are the consequences of cutting all funding for 
programs that recruit and retain diverse teachers?
    Ms. Mote. Well, I think every member of this Committee can 
say that we need to keep teachers at the center of that 
education enterprise. When we do not have teachers to serve the 
communities that they are from, when we do not have classrooms 
that are reflective of the communities that students are coming 
from, we know actually that that affects student outcomes.
    Research shows us that having a connection with an adult in 
school, not only as I shared before, reduces suicidal ideation, 
but it is the single greatest determinant of whether or not 
student learning outcomes will excel.
    Ms. Adams. Let me move on and talk about the data. Good 
policy depends on good information, but last month the Trump 
administration eliminated nearly 900 million dollars in 
contracts from the Institute of Education Sciences, the 
research arm of the Department of Education.
    That is the same office that helps districts and 
policymakers understand what is working and what is not. Ms. 
Mote, how does cutting IES funding impact the school's ability 
to use AI responsibly and equitably? Could that data gap leave 
underfunded schools even further behind?
    Ms. Mote. Yes. If we do not know what works for whom, and 
under what conditions, how can we choose the right tools for 
students from all backgrounds?
    Ms. Adams. Thank you, Ms. Mote. If we want students to 
succeed in the age of AI, we cannot just chase shiny new tools, 
we have to invest in the people, and the research and the 
infrastructure that make innovation safe and meaningful, and I 
thank you all very much. Mr. Chairman, I am going to yield back 
those few other minutes that I have. Thank you.
    Chairman Kiley. The Representative from Pennsylvania, Mr. 
Mackenzie, is recognized for 5 minutes.
    Mr. Mackenzie. Thank you to the Chairman, and I appreciate 
all of the testimony being provided today, very important 
topic. Integrating AI into our education system has huge 
potential benefit for our students. I think we are all aware 
of, even right now in its early stages, what it can do to 
increase educational opportunity and customization for 
students, so a lot of benefit there.
    We are also talking about some of the elements that we need 
to be aware of to make sure that we protect privacy and 
security for our students, great topics, but earlier we heard 
from a number of testifiers about bias that can be built into 
the AI systems that are being incorporated into the classroom.
    I think we are all aware that depending on who nurtures or 
trains the AI, they are going to end up putting out different 
results at the end of the day. Ultimately, I think what we want 
for our students is a maximally seeking truth AI. We want the 
truth to be provided to our students, so it should not be a 
corporation. It should not be a foreign entity that is maybe 
weighing in on the results that are coming from AI.
    It should not even be a politician on this panel saying 
that we need to reweigh results, so that we get the accurate 
outcome that we want for our students. How do we, and how 
should everybody be thinking about what AI we are selecting? 
What actual critiques or reviews of that AI is being made 
available publicly when we see that bias or misinformation is 
being included in results?
    How does that get communicated to the decisionmakers in 
school districts or local entities that are actually picking 
the AI systems? I will go first to Dr. Rafal-Baer.
    Ms. Rafal-Baer. Thank you so much, and incredibly 
thoughtful. I think there is a couple of layers to that answer. 
I think from our perspective we think states having AI 
assurance labs is one place to be able to examine this. Having 
one place where the public has a level of transparency about 
how the model on the front end before it is approved for use, 
what it has gone through to be rigorously tested.
    Then in an ongoing way, looking at the outputs of the 
model, knowing that these models evolve overtime as people are 
using them. I think you are also hitting a larger topic that we 
have talked about a few times, about AGI. What happens when we 
get to a place where we are working alongside models that are 
doing things at the same level of humans, knowing that there 
are levels of bias within that?
    I think that is where the importance of critical thinking, 
and helping to really infuse in our students, understanding 
about ethics, and ethical considerations is such a critical 
role of our K-12 system. To be able to make sure that we are 
bringing those conversations often into these community 
engagement sessions in thinking about this.
    We will never get rid of bias, but we can help to make sure 
we are shaping a generation of learners who deeply understand 
the power and importance of human relationship, and when and 
where to use AI and under what conditions, and how to evaluate 
those outputs consistently.
    Mr. Mackenzie. I will go to anybody else on the panel. Do 
you guys have any thoughts or are there resources available 
where people can go and see which AI systems are actually 
providing on a whole, collectively, maximally truth-seeking 
information? I think this is going to be a real challenge for 
everybody that is adopting this to have this kind of decision 
made in a classroom.
    It is akin to selecting your curriculum, or your textbooks. 
At the same time there is a lot less information out there 
right now about these AI systems. How do you as educators see 
that?
    Mr. Dobrin. I think that is a fantastic question, and I 
want to reinforce what Dr. Rafal-Baer has said. I think there 
is a step back in asking that question. I think making the 
analogue to textbooks is very important because we are going to 
have to think about adoption in the classroom of various 
platforms in the same way we think about curricular options of 
textbooks, both in finance models and in content models.
    I think an answer to your bigger question, this is where 
the importance of demystifying these tools becomes critical. If 
we are going to move into the critical thinking position about 
having students be able to evaluate the information that is put 
before them, they first have to understand how those systems 
work. What an LLM does, where the data is pulled from, so that 
they can begin those evaluative processes. In other words, by 
demystifying the function of a generative AI tech, allows the 
student then to see the working model and understand these are 
now the questions I have to ask.
    As to resources for identifying which platforms are more 
viable than others, the first thing we have to acknowledge is 
that what we need is more research and development for 
platforms specifically designed for education. We do a lot in 
education of retrofitting to technologies that have already 
existed and put in play in industry and other places.
    We actually need platforms that are designed for education 
specifically. This is why Mr. Chism's approach to having his 
own independent server and his own model is a very kind of 
forward-thinking model. I think the demystification process is 
critical.
    Mr. Mackenzie. Well, I will wrap up. We are over time, but 
I want to thank all of you again for your very thoughtful input 
today. I think this is again, an incredibly important topic as 
we look to advance education in our country, and I appreciate 
particularly that last point about demystifying the systems 
because I'll date myself very quickly here.
    When I was growing up, scientific calculators were a new 
invention, but at the same time we were always trained to 
actually understand the fundamentals and the processes before 
we used the scientific calculators to get the answer in a 
quicker fashion. Thank you again, and I yield back.
    Chairman Kiley. The Representative from Illinois, Ms. 
Miller is recognized for 5 minutes.
    Mrs. Miller. Thank you and thank you to all the witnesses 
who have come. This is a very interesting conversation. Mr. 
Chism, we have heard a lot about the potential benefits of AI 
in schools, but one of the longstanding priorities of this 
Committee has been student privacy.
    There is a lot that people do not understand, or were 
rightly concerned about when it comes to AI's impact on student 
privacy. You mentioned this a bit in your testimony, but what 
student privacy pitfalls do you think districts should be aware 
of, and how has your district handled this?
    Mr. Chism. That is a great question, and again, student 
privacy, we have to file under FERPA law, so we have got to 
make sure we are in compliance with that. It is real simple. I 
do not want to make this--over simplify this, but it really is 
simple.
    If you are using a public entity, I do not care if it is 
Open AI, Grok, any of these others, you cannot put student 
identifiable information on it, or teacher identifiable 
information. Really teaching people to scrub files, that is the 
biggest thing if you are going to use an Excel file that has 
this information, knowing what you need to take out.
    That is a way that we handle that if you are using 
something like Open AI. However, we alleviate that by using our 
own server. I do not have to worry about FERPA laws because it 
does not face the internet in any way, shape or form. Again, we 
have done trainings with teachers and administrators. We can 
put any file that we want on our own server, and we do not ever 
have to worry about student privacy there because it is housed 
on our own campuses, so we do not have to worry about that.
    I think that is a way to do that. Again, that is a pretty 
expensive way to do it. However, on the flip side, if you are 
using this as a district, such as with Open AI, you are 
eventually going to start having to pay for tokens. In the end 
we, as a district, have decided that it is going to be more 
cost-effective to just go ahead and pay for our own server, 
than try to pay for tokens with one of those as well.
    I hope that answers your question. That is the crux of what 
we do. It is good training first, if you are using a public 
server. Then on the other side, we just create our own, so we 
alleviate the problem altogether.
    Mrs. Miller. Thank you. I appreciate how you are leading 
the way there. Mr. Chism, I have another question. One of the 
most frequent concerns we hear about AI in schools is that 
students will use AI as a crutch. For example, students will 
ask AI to write their essays or answer their homework 
questions.
    I know there are limits to how much a school district can 
do, especially once a student goes home for the day, but how 
are you thinking through these challenges?
    Mr. Chism. That is probably my favorite question. Really, 
we have to rethink what we do in education. I told our teachers 
in all of the presentations that I do that we have got to 
rethink what we do in the classroom. Basically, we have to 
control the 8 hours that we have.
    We cannot worry about other things outside of school 
because understand, they are going to use it, so really, we 
need a heavier focus on what we do in the classroom and control 
the 8 hours that we have. That means heavier weighting on 
grades for the things that we do in classrooms.
    It means that again, we have to shift, instead of sending 
things home for kids to write, we need to take care of those 
things within the classroom, and worry about the controllables, 
and again, we cannot worry about what happens outside. We just 
have to shift our focus to things like having projects done in 
class, or having students explain their work.
    Why do I care where the information came from if the 
student can actually give that information back and teach 
someone else? They know the material. That is the ultimate 
thing that we are trying to do within the district, so control 
the 8 hours that we have.
    Mrs. Miller. Thank you. Dr. Dobrin, one especially exciting 
aspect of the AI adoption is the potential that this technology 
can transform education for special needs students, and I have 
two grandsons with special needs. For instance, AI powered 
programs can facilitate speech detects software, or help 
visually impaired students participate more fully in class. Can 
you talk more about how AI is being used to practically enhance 
special ed?
    Mr. Dobrin. Yes, thank you for that question. The first 
thing I need to say is that I do not have experience in special 
education specifically, and Ms. Mote earlier had addressed some 
of this, and so I defer to her on that.
    What you are talking about specifically are assistive 
technologies. For me, in terms of special needs students, and 
also really for any student, one of the most enriching 
possibilities that we are seeing with AI right now is in 
customized learning pathways. The ability for a student to move 
through information and content, and all the things associated 
with education in a way that is more effective for that 
learner.
    Now, I tend to also put a stop gap on where that becomes 
beneficial once we get past higher ed into credentialing, but I 
think that customized learning pathways right now are one of 
the most important things we are developing, particularly in 
terms of real time assessment, and the ability for students in 
that moment of assessment to adjust their learning path to see 
I did not understand this concept. I now have this other way of 
learning about it.
    That to me is one of the most important things we are 
doing.
    Mrs. Miller. Thank you and I yield back.
    Chairman Kiley. I will now recognize the Ranking Member of 
the Full Committee, Mr. Scott.
    Mr. Scott. Thank you, Mr. Chairman. Ms. Mote, I think it is 
clear that there are some benefits and some concerns about AI. 
Some of the benefits--is there any question that AI can expand 
the capability of a classroom teacher? If so, are there other 
areas where it is particularly good or bad--language, math, 
science?
    Ms. Mote. Well, I am under the idea that AI literacy is a 
foundational literacy, and frankly, AI itself is the best 
revenge of English teachers because it is all about how do you 
ask good questions? When we think about how to build literacy 
across our disciplines, if we really want to realize the 
productivity gains that Mr. Chism is talking about, we want to 
think through ways that we can support students with 
disabilities, particularly around early screeners for dyslexia.
    I am the mom of a dyslexic student. It would have been so 
much more valuable if I could have had my son screened earlier, 
so that I could have intervened earlier. AI is allowing 
teachers to take that practice that used to take weeks into an 
afternoon, as long as they keep a human in the loop.
    For me, building that competency, across all disciplines, 
is what we need to meet this moment, Mr. Scott.
    Mr. Scott. You have cautioned about leaving out human 
interaction. Do you have to choose between AI and human 
interaction?
    Ms. Mote. No. Actually, that is not what AI models want you 
to do. AI models want you to put your human interactions into 
them, so they get smarter and can be trained faster. When we 
talk about intelligent tutors adapting to students' language, 
or identifying dialectical differences, it is because the AI is 
actually trying to learn from the interaction that it is having 
with a person.
    Mr. Scott. Now, can the AI help with assessing the progress 
of a student in the classroom?
    Ms. Mote. This is a place I think we need to be cautious. I 
think we always need to have a human in the loop, particularly 
when it comes to assessment. Tools have shown that they 
disproportionately flag students for cheating when they have 
not cheated by race or by dialectical difference, for example, 
our rural students.
    I think we need to be cautious when we think about 
assessments, and make sure that we are engaging humans.
    Mr. Scott. One of the access questions is obviously cost. 
How much does all of this cost if you are integrating AI into 
the classroom? Is it very expensive?
    Ms. Mote. Well, I think Mr. Chism talked a little bit about 
the costs that he has taken on in terms of buying a server and 
buying technology. It is not just those hardware costs, Mr. 
Scott, it is the training that you need to do. It is the 
education. I think we are in a situation right now where 
whether you have access to an intelligent tutor like Khanmigo 
is about $99.00 a student.
    I know that that is a decision some school districts cannot 
make right now because they are constrained when it comes to 
resources and support. It is not just about the tool it is 
about the training that has to go alongside of it.
    Mr. Scott. I guess getting benefits to everyone on an 
equitable basis, what are some of the other barriers?
    Ms. Mote. Yes. I think when we think about AI's use in 
education, we are going to have to really think about a 
partnership. A partnership with families and communities. A 
partnership with students, and frankly a partnership with 
industry.
    Industry is trying right now to do work to make their tools 
more safe, more effective, more transparent, more reliable, but 
they need help in terms of guidelines and guardrails. If 
industry has to conform to 50 different states privacy laws, 
data laws, or AI laws, imagine what that means for an economy 
of scale, or equitable access to tools.
    If we are going to ask everybody to build an AI tool for 
the State of Delaware, the State of Virginia, or the State of 
Colorado, we miss the opportunity frankly, to be able to 
appreciate the richness and diversity that we have in the 
United States.
    Mr. Scott. Bias has been mentioned. There are a number of 
ways that bias can pop up in AI. Can you talk about that in the 
remaining time?
    Ms. Mote. In 30 seconds, OK. I will just give one example. 
You know, I think as I said, AI looks for the interactions that 
it has with humans, and so there are some really great research 
studies that have been done that when an AI tool is trained, 
for example, on just assessment, it does not flag students more 
or less from one background.
    If somebody uses a word like y'all, which I use all the 
time, it flags that student potentially for cheating. 
Dialectical difference, not just race or socioeconomic status, 
are things we need to be paying attention to.
    Mr. Scott. Thank you, Mr. Chairman.
    Chairman Kiley. The Representative from Guam, Mr. Moylan, 
is recognized for 5 minutes.
    Mr. Moylan. Thank you, Mr. Chairman. Emerging AI 
technologies have the potential to enhance our quality of life 
in many ways, especially in increasing efficiency and learning 
outcomes for students. AI may also provide innovative support 
for students who need individualized learning plans, especially 
for special education.
    However, we must be aware of the risk of these 
technologies, especially with privacy concerns. Proper safety 
protocols can mitigate many of these risks, and best practice 
guidance are being constantly issued to mitigate these risks. 
As we enter the future economy, preparing our youth to enter 
the workforce where employers are increasingly interested in AI 
into their business models, it is especially important.
    Many paths toward high-earning jobs exist. AI and computer 
science can offer alternative paths toward high-skilled 
employment, where many opportunities do not necessarily require 
a traditional 4-year degree to participate. For my first 
question, Dr. Rafal-Baer, you mentioned that implementing AI 
should reflect the priorities of each community.
    As employers begin adopting more and more AI tools into 
their business, it is critical that we ensure our students are 
ready to join the modern workforce. Can you talk specifically 
about what kind of AI skills employers are looking for?
    Ms. Rafal-Baer. Yes, thank you. I think we think about this 
generation who is going to be moving on. The thing that they 
are going to be most critically asked about is just their basic 
foundations around AI, understanding what it is, how it works, 
its limitations, its risks, and being able to communicate about 
that.
    I think what we are going to see over time is much more of 
a desire for our education setting to take on more of the AI 
literacy work, to evolve from what has been digital literacy, 
which we are now well past, and really starting to understand 
how to work alongside AI, what use cases are appropriate, and 
then how to always think about the relationship with other 
humans.
    We know how important it is to be able to have 
relationships to grow high growth, high-impact teams, and 
increasingly those kinds of skills are going to have to be ones 
that we are reinforcing in an AI age. The last thing I would 
say I think in all of this it is going to be just a huge 
undergrounding feeling around ethics.
    I think employers are going to want to know that students 
are graduating with a keen understanding about ethical 
considerations. That is something that gets infused throughout 
the curriculum. It is not one subject area's domain, but it is 
why deep stakeholder engagement matters so much, to surface the 
kinds of ethical questions that are being asked and then use 
that in real world ways in the classroom.
    Mr. Moylan. Thank you. For Dr. Dobrin, in your testimony 
you mentioned that AI literacy is crucial for students entering 
the workforce. Teaching AI may come with many challenges, and 
education leaders might now know the best way to begin.
    What advice would you give our teachers and education 
officials that want to ensure that students are equipped with 
skills employers need?
    Mr. Dobrin. Thank you, yes. The first place I would begin, 
and it has been mentioned several times, is this concept of AI 
literacy, which I think we actually, and I have written about 
this. We need to expand into AI literacy, AI competencies, and 
AI fluencies, because different work paths require different 
levels of skillsets specifically from AI.
    In the work that I do with industry, usually the No. 1 
topic, and it has come up here several times that employers are 
looking for is data security, particularly when they are 
working with their own proprietary datasets. My advice, very 
specifically, and I work with lots of educators about this, is 
that we adopt a model that we have used in the United States 
for a long time that I call industry to curriculum.
    What that means is understanding where industry--a specific 
industry, is deploying AI and how, and then adjusting our 
curriculum to that. We can see this in a lot of the disciplines 
we teach, particularly in higher ed, nursing, law, engineering, 
these are disciplines that are designed to provide credentials 
toward specific industries.
    Once the industries start determining what exactly they are 
looking for in terms of AI skills, then we adopt curriculum to 
that as a way to move education into workplace development.
    Mr. Moylan. Excellent, thank you. Last question, Dr. Rafal-
Baer, your professional background is in special education. Can 
you speak more about how educators can leverage AI tools to 
tailor plans for students enrolled in special ed?
    Ms. Rafal-Baer. I think one of the most exciting places on 
this is both the ability for educators and caregivers to work 
together, and to use AI to help to simplify what can sometimes 
feel like a lot of jargon and technical language and can make 
it difficult to best support students.
    Then for educators, being able to effectively communicate 
and provide tips and ways in which families can work at home to 
best support the students, and over time, and provide real 
pathways.
    Mr. Moylan. Thank you. Thank you, Mr. Chairman, thank you 
panel.
    Chairman Kiley. Thank you very much to all of our 
witnesses, and I will now recognize Ranking Member Bonamici for 
a closing statement.
    Ms. Bonamici. Thank you, Mr., Chairman, but thank you 
especially to the witnesses for traveling here for your 
testimony and brining your expertise on this excellent panel. 
As we conclude, one thing is abundantly clear, artificial 
intelligence will play a significant role in shaping the future 
of education, and we should be breaking down, rather than 
creating barriers for the students and the educators who are 
navigating this transformation.
    AI and education presents tremendous opportunities, but 
also serious challenges as we heard. If implemented 
responsibly, it has the potential to enhance learning, provide 
personalized instruction, equip students with the skills they 
need to thrive in an increasingly digital world.
    Without guardrails, we will see decreasing access to 
opportunity, a deepening of inequities we are trying to 
overcome, and risk serious privacy and security breaches. The 
Department of Education has been instrumental in closing 
opportunity gaps, providing students in low-income and rural 
communities, and students with disabilities with access to the 
resources they need.
    Stripping away these protections, whether they be through 
the dismantling with the Department of Education's Office of 
Educational Technologies, or cuts to Title I, or IDEA, or the 
weakening of the Office of Civil Rights. These would all be a 
grave mistake.
    The Federal Government would be leaving schools to navigate 
these challenges, exacerbating disparities and failing the very 
students who rely most on Federal support, as we have heard 
particularly low-income students, and students with 
disabilities.
    We have seen this play out before when the digital divide 
threatened to leave millions of students behind in the early 
days of the internet. It was Federal leadership that helped 
bridge the gap. We cannot repeat history by allowing AI to 
widen existing inequities.
    Instead, we must guarantee that every student, no matter 
their ZIP Code, has access to the tools and knowledge they need 
to succeed in the classrooms and careers of tomorrow. I know 
this is the K-12 Subcommittee, but as also a member of the 
Higher Education Committee we talked about workforce, but one 
of the things I do want to mention, and I think our witnesses 
made a great case for it today when we got a head of an English 
department and someone who is talking a lot about ethics.
    We have got Dr. Dobrin and Dr. Rafal-Baer. Tech executives 
I have heard recently say we do not need just computer science 
majors and people who code. We need people who are philosophers 
and ethicists as well, as we navigate these serious challenges. 
The stakes could not be higher. We cannot afford to get this 
wrong. I urge my colleagues to stand with us and defend public 
education, protect vital resources, and provide every student 
with a fair shot at success, and I look forward to the 
continued conversation. Thank you again, Mr. Chairman, and I 
yield back.
    Chairman Kiley. Thanks very much. I think this has been a 
great hearing. I think that we have moved the ball forward, 
improving a bit our institutional understanding of a phenomenon 
that is really beyond any of our full understanding, but that 
we urgently need to grapple with, nonetheless.
    Ms. Mote likened AI to electricity and being what she 
called an arrival technology. Ms. Rafal-Baer noted that the 
arrival of AGI, artificial general intelligence, could be 
coming sooner than has been thought. This is defined 
differently by different people, who generally refers to a 
system that develops capabilities that exceed human 
capabilities across any domain.
    The conversation about this has been largely limited to the 
leading labs and those who follow them, but it is something 
that needs to--that we all need to be part of throughout the 
country, given the profound impacts it is going to have on all 
of our lives.
    Today specifically, we have gotten a window into how the 
possibilities offered by AI and education. I think that the 
main takeaway today is that those possibilities really are 
boundless. Mr. Chism noted that education is often the last 
change when it comes to technology, or perhaps really anything 
else, but that really cannot be the case here.
    I think we already know enough to know that those schools, 
those districts, those states that adopt and integrate AI in 
the right ways are going to do very well by their students. 
They are going to propel student success and are going to 
outperform those jurisdictions that do not.
    By the way, in other countries like China, has been very 
focused on this topic, and so our global competitiveness is at 
stake as well. Of course, in order to do this effectively, we 
do need to have professional development, which is why I think 
what Mr. Chism has been doing throughout Mississippi is 
fantastic, that supports teachers.
    We learned today that most teachers still are not getting 
training when it comes to integrating AI. I agree with Mr. 
Mannion fully, who said that AI can support and not supplant 
teachers.
    In fact, we had a great story from Mr. Chism about a 
teacher who was going to retire because of all the hours they 
put in in grading, and that AI was able to entice her to stay 
in the classroom because it allowed her to focus on the aspects 
of teaching that she loved, and that made the biggest 
difference.
    Another thing I think we heard today that I agree with, is 
that there is a pressing need for more research when it comes 
to how AI is being deployed, and the ways in which it can be 
deployed most effectively, what modalities are most effective, 
what parts of the traditional lesson plan it is best 
incorporated into.
    As in all things education, it is best when we have 
measurable outcomes, and then can let that guy, our pedagogy 
going forward. Ms. Rafal-Baer has noted about everything going 
on at the State level in terms of research. I think that there 
perhaps is a Federal rule here as well, at least for the sort 
of thing we are doing today for coordination, for sharing best 
practices.
    I think Ms. Rafal-Baer mentioned having a summit of some 
kind, which I think is a great idea. Then you also mentioned, 
Ms. Rafal-Baer, the importance of local engagement as well with 
stakeholders, with parents, with the community. If the question 
of efficacy that research can answer, but there is also the 
question of values, making sure that these technologies are 
being integrated in classrooms in ways that are constant with 
the values of the surrounding community.
    This is actually a topic that is broadly applicable to AI 
in many facets of society. Actually, in 2018, I was a member of 
the California legislature. I sponsored a resolution about the 
benefits of AI, but how we needed to make sure it was 
consistent with human values as the adoption spread.
    Education provides a very important example of that. We 
have talked today about Ms. Mote, which mentioned the idea of 
keeping humans in the loop, which I think is a useful concept, 
and how there really still is no replacement for human 
interaction in many domains in society, and many parts of our 
lives, and that is probably true most of all when it comes to 
education, given the formative nature of education for young 
people.
    We also talked about values related to data and privacy, 
which Ms. Miller, among others mentioned, and then sort of both 
sides mentioned different takes on the question of bias, and 
how we need to make sure that there are safeguards in place, so 
that there is not bias incorporated, as we use AI tools.
    Finally, I think that, you know, perhaps the biggest 
takeaway from the specific perspective of education policy 
today is the enormous potential that AI has to expand 
opportunity to close achievement gaps, to democratize access to 
the very best instruction.
    We, for too long in this country, have had these terrible 
achievement gaps where opportunity is a product of where you 
live, and where you grow up. In my view, that is a result of 
poor policy choices that have been made over the course of many 
years and decades. Perhaps not all of my colleagues will agree 
with me on that, but I would hope that we could find some 
common ground on this issue.
    That we should all be excited about tools that allow for 
instruction to be tailored and in an individualized way. 
Several people, including Ms. Miller, Mr. Thompson, mentioned 
the unique potential to assist students with special needs in 
this area, and the ability to have this individualized 
instruction that is, you know, mindful of a student's strengths 
and weaknesses, the prior knowledge that they bring to bear, 
their interests even, has extraordinary potential.
    I want to thank Mr. Dobrin as well for his discussion of 
how all of this can also democratize access to the sort of 
skills that are successful--that are needed for success in the 
workplace.
    Thank you again everyone. Thank you for some who endured 
long travels to be here, and I think it was a great hearing, 
and I look forward to continuing this conversation going 
forward. The Subcommittee stands adjourned.
    [Whereupon, at 12:21 p.m., the Subcommittee was adjourned.]

    [Questions and responses submitted for the record by Ms. 
Erin Mote follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]

                       [all]