[Senate Report 117-40]
[From the U.S. Government Publishing Office]
Calendar No. 142
117th Congress } { Report
SENATE
1st Session } { 117-40
_______________________________________________________________________
ARTIFICIAL INTELLIGENCE TRAINING FOR THE ACQUISITION WORKFORCE ACT
__________
R E P O R T
of the
COMMITTEE ON HOMELAND SECURITY AND
GOVERNMENTAL AFFAIRS
UNITED STATES SENATE
to accompany
S. 2551
TO REQUIRE THE DIRECTOR OF THE OFFICE OF
MANAGEMENT AND BUDGET TO ESTABLISH OR OTHERWISE
PROVIDE AN ARTIFICIAL INTELLIGENCE TRAINING PROGRAM
FOR THE ACQUISITION WORKFORCE, AND FOR OTHER PURPOSES
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
October 5, 2021.--Ordered to be printed
__________
U.S. GOVERNMENT PUBLISHING OFFICE
29-010 WASHINGTON : 2021
COMMITTEE ON HOMELAND SECURITY AND GOVERNMENTAL AFFAIRS
GARY C. PETERS, Michigan, Chairman
THOMAS R. CARPER, Delaware ROB PORTMAN, Ohio
MAGGIE HASSAN, New Hampshire RON JOHNSON, Wisconsin
KYRSTEN SINEMA, Arizona RAND PAUL, Kentucky
JACKY ROSEN, Nevada JAMES LANKFORD, Oklahoma
ALEX PADILLA, California MITT ROMNEY, Utah
JON OSSOFF, Georgia RICK SCOTT, Florida
JOSH HAWLEY, Missouri
David M. Weinberg, Staff Director
Zachary I. Schram, Chief Counsel
Michelle M. Benecke, Senior Counsel
Pamela Thiessen, Minority Staff Director
Andrew C. Dockham, Minority Chief Counsel and Deputy Staff Director
Sam J. Mulopulos, Minority Professional Staff Member
Laura W. Kilbride, Chief Clerk
Calendar No. 142
117th Congress } { Report
SENATE
1st Session } { 117-40
======================================================================
ARTIFICIAL INTELLIGENCE TRAINING FOR THE ACQUISITION WORKFORCE ACT
_______
October 5, 2021.--Ordered to be printed
_______
Mr. Peters, from the Committee on Homeland Security and Governmental
Affairs, submitted the following
R E P O R T
[To accompany S. 2551]
[Including cost estimate of the Congressional Budget Office]
The Committee on Homeland Security and Governmental
Affairs, to which was referred the bill (S. 2551) to require
the Director of the Office of Management and Budget to
establish or otherwise provide an artificial intelligence
training program for the acquisition workforce, and for other
purposes, having considered the same, reports favorably thereon
without amendment and recommends that the bill do pass.
CONTENTS
Page
I. Purpose and Summary..............................................1
II. Background and Need for the Legislation..........................2
III. Legislative History..............................................3
IV. Section-by-Section Analysis of the Bill, as Reported.............3
V. Evaluation of Regulatory Impact..................................4
VI. Congressional Budget Office Cost Estimate........................4
VII. Changes in Existing Law Made by the Bill, as Reported............5
I. PURPOSE AND SUMMARY
S. 2551, the Artificial Intelligence Training for the
Workforce Act, or the ``AI Training Act,'' requires the Office
of Management and Budget (OMB) to establish or otherwise
provide a training program on artificial intelligence (AI) for
the federal acquisition workforce, in order to ensure that the
workforce has knowledge of the capabilities and risks
associated with AI. The bill defines artificial intelligence
and the covered workforce members targeted for the training,
including program and procurement personnel, and others as
designated by the head of the executive agency to participate.
The Director of OMB, in coordination with the Administrator of
General Services and others determined relevant by the
Director, must develop and provide the training program to the
covered workforce not later than one year after the date of
enactment.
Additionally, S. 2551 outlines topics to be covered in the
AI training program, including introductory concepts; the
science underlying AI and how it works; benefits of AI for the
federal government; the risks posed by AI and ways to mitigate
these risks to ensure it is reliable, safe and trustworthy; and
future trends in AI, including trends for homeland and national
security and innovation. The bill requires the training program
to be updated not less frequently than every two years, in
order to incorporate new information and ensure the program
continues to satisfy requirements. An interactive format for
the training involving technologists, scholars, and other
experts is encouraged, and metrics on participation and
feedback must be maintained. The bill contains a sunset
provision effective ten years after the date of enactment.
II. BACKGROUND AND NEED FOR THE LEGISLATION
As the federal government ramps up investment in AI,
employees who acquire and manage this technology need basic
training on its capabilities and risks. The need for workforce
training on AI has been recognized by experts such as the
National Security Commission on Artificial Intelligence
(NSCAI).\1\ NSCAI is particularly concerned with the
implications of an under-trained federal acquisition workforce,
noting: ``Government agencies that rely solely on contractors
for digital expertise will become incapable of understanding
the underlying technology well enough to make successful
acquisition decisions independent of contractors.''\2\ The AI
Training Act addresses the need to develop an acquisition
workforce that is prepared to engage with the technical and
ethical implications of AI as the United States seeks to
maintain a global advantage with this technology.
---------------------------------------------------------------------------
\1\National Security Commission on Artificial Intelligence, First
Quarter Recommendations (Mar. 2020) (https://www.nscai.gov/wp-content/
uploads/2021/01/Summary-of-NSCAI-Q1-Recommendations.pdf).
\2\National Security Commission on Artificial Intelligence, Final
Report (Mar. 2021) (https://www.nscai.gov/wp-content/uploads/2021/03/
Full-Report-Digital-1.pdf).
---------------------------------------------------------------------------
Although multiple executive orders have promoted the use of
AI across the federal government,\3\ there has not been a
concerted effort to train the civilian workforce that
identifies, develops, buys, and manages AI capabilities to
support the mission. While informal groups exist, such as the
General Services Administration's AI Community of Practice,\4\
they are no substitute for a structured and ongoing program
that ensures proper government adoption of AI. Market research
predicts that the federal government will spend nearly $1
billion on AI by 2023, which will greatly increase productivity
and improve engagement with the public if the federal workforce
learns to use it effectively.\5\
---------------------------------------------------------------------------
\3\Exec. Order No. 13960, 85 Fed. Reg. 78939 (Dec. 3, 2020); see
also Exec. Order No. 13859, 84 Fed. Re. 3967 (Feb. 11, 2019).
\4\1A Digital.gov, Artificial Intelligence Community (https://
digital.gov/communities/artificial-intelligence/) (accessed Sept. 24,
2021).
\5\Nextgov, The Case to Increase Workforce Training for Artificial
Intelligence (Aug. 19, 2020) (https://www.nextgov.com/ideas/2020/08/
case-increase-workforce-training-artificial-intelligence/167782/).
---------------------------------------------------------------------------
Along with positive impacts for agencies and society, there
are parallel causes of concern that AI can, if improperly or
maliciously applied, result in invasions of privacy and have
discriminatory effects.\6\ The ethical use of AI is an
especially urgent issue.\7\
---------------------------------------------------------------------------
\6\Brookings Institution Report, Protecting Privacy in an AI-driven
World (Feb. 10, 2020) (https://www.brookings.edu/research/protecting-
privacy-in-an-ai-driven-world/).
\7\Government Accountability Office, Artificial Intelligence: An
Accountability Framework for Federal Agencies and Other Entities (GAO-
21-519SP) (Jun. 30, 2021); Government Accountability Office, Facial
Recognition Technology: Federal Law Enforcement Agencies Should Better
Assess Privacy and Other Risks (GAO-21-518) (Jun. 29, 2021); see also
Forensic Technology: Algorithms Strengthen Forensic Analysis, but
Several Factors Can Affect Outcomes (GAO-21-435SP) (Jul. 6, 2021).
---------------------------------------------------------------------------
III. LEGISLATIVE HISTORY
Chairman Gary Peters (D-MI) introduced S. 2551, the AI
Training Act, on July 29, 2021, with Senator Portman (R-OH).
The bill was referred to the Senate Committee on Homeland
Security and Governmental Affairs. The Committee considered S.
2551 at a business meeting on August 4, 2021. The Committee
ordered the bill reported favorably without amendment by voice
vote en bloc. Senators present for the vote were: Peters,
Carper, Hassan, Sinema, Rosen, Padilla, Ossoff, Portman,
Johnson, Lankford, Romney, Scott, and Hawley.
IV. SECTION-BY-SECTION ANALYSIS OF THE BILL, AS REPORTED
Section 1. Short title
This section provides that the bill may be cited as the
``AI Training Act.''
Section 2. Artificial intelligence training programs
Subsection (a) defines ``artificial intelligence,'' ``AI
training program,'' ``covered workforce,'' ``Director,'' and
``executive agency.''
Subsection (b)(1) requires the Director of the Office of
Management and Budget, in coordination with the Administrator
of General Services and others, to develop and implement or
otherwise provide an AI training program for the covered
workforce.
Subsection (b)(2) states the purpose of the AI training
program as ensuring that the covered workforce has knowledge of
the capabilities and risks associated with AI.
Subsection (b)(3) prescribes information to be covered in
the AI training program, including the science underlying AI
and how it works; introductory concepts related to AI; ways in
which AI can benefit the federal government; the risks posed by
AI, including discrimination and risks to privacy; ways to
mitigate these risks and efforts to create and identify AI that
is reliable, safe, and trustworthy; and future trends in AI,
including trends for homeland and national security and
innovation.
Subsection (b)(4) requires the Director of OMB to update
the training program not less frequently than once every two
years in order to incorporate new information and ensure that
the AI training program continues to satisfy the requirements
of (b)(3).
Subsection (b)(5) encourages the Director of OMB to include
interactive learning in the AI training program involving
technologists, scholars, and other private, public and
nonprofit sector experts.
Subsection (b)(6) requires the Director of OMB to track
metrics on participation and feedback related to the AI
training program.
Subsection (b)(7) holds that Section 2 shall cease to have
effect 10 years after enactment.
V. EVALUATION OF REGULATORY IMPACT
Pursuant to the requirements of paragraph 11(b) of rule
XXVI of the Standing Rules of the Senate, the Committee has
considered the regulatory impact of this bill and determined
that the bill will have no regulatory impact within the meaning
of the rules. The Committee agrees with the Congressional
Budget Office's statement that the bill contains no
intergovernmental or private-sector mandates as defined in the
Unfunded Mandates Reform Act (UMRA) and would impose no costs
on state, local, or tribal governments.
VI. CONGRESSIONAL BUDGET OFFICE COST ESTIMATE
U.S. Congress,
Congressional Budget Office,
Washington, DC, September 28, 2021.
Hon. Gary C. Peters,
Chairman, Committee on Homeland Security and Governmental Affairs, U.S.
Senate, Washington, DC.
Dear Mr. Chairman: The Congressional Budget Office has
prepared the enclosed cost estimate for S. 2551, the Artificial
Intelligence Training for the Acquisition Workforce Act.
If you wish further details on this estimate, we will be
pleased to provide them. The CBO staff contact is Aldo
Prosperi.
Sincerely,
Phillip L. Swagel,
Director.
Enclosure.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Artificial Intelligence (AI) allows computer systems to
perform tasks that require human intelligence. S. 2551 would
require the Office of Management and Budget (OMB) to establish
a training program to provide federal procurement officials
with an introductory understanding of the operational benefits
and privacy risks of AI.
Using information from OMB and other agencies about current
efforts to use AI across the federal government, CBO
anticipates that OMB would need five full-time employees to
create and manage the training program. CBO estimates that
staff salaries and technology costs to deliver the training
would total $2 million annually. Accounting for the time needed
to hire new employees and develop the training, CBO estimates
that implementing the bill would cost $9 million over the 2021-
2026 period; such spending would be subject to the availability
of appropriations. For this estimate, CBO assumes that the bill
will be enacted in fiscal year 2022.
The CBO staff contact for this estimate is Aldo Prosperi.
The estimate was reviewed by Leo Lex, Deputy Director of Budget
Analysis.
VII. CHANGES IN EXISTING LAW MADE BY THE BILL, AS REPORTED
This legislation would make no change in existing law,
within the meaning of clauses (a) and (b) of subparagraph 12 of
rule XXVI of the Standing Rules of the Senate, because this
legislation would not repeal or amend any provision of current
law.
[all]