AI TASK FORCE – Crunch

AI Task Force formed during the second session of the 118th Congress, to address the challenges and
opportunities surrounding Artificial Intelligence (AI), consisting of bipartisan members and aims to
guide Congress in crafting policies that foster responsible AI development and deployment while
maintaining the United States' leadership position in this transformative technology. The Task Force
consulted with a wide range of experts, including business leaders, government officials, technical
experts, academics, and legal scholars.

The ultimate goal is to ensure that AI's benefits—ranging from improvements in quality of life,
health, jobs, security, and economic prosperity—are maximized for Americans and globally. By
taking an agile approach, the Task Force envisions that the U.S. can continue to lead AI innovation
while promoting responsible practices that will help guide the global conversation on AI policy.

Principles for AI Policy Framework

Iden Assess whether AI issues are truly new, significantly changed by AI, or whether existing policies are adequate.
tify New due to AI capabilities
AI Changed by AI impact
Issu Existing issue, not changed by AI
e
Nov
elty
Pro
mot
Encourage responsible AI development to maintain U.S. leadership and harness AI's potential for economic growth and
e AI societal benefits.
Inno
vati
Prot
on
ect Identify and mitigate risks, both accidental and malicious, while fostering innovation with a thoughtful, risk-based
Agai approach.
nst
AI Address potential harms
Risk Leverage AI for solutions to its own challenges
s
and
Emp
Har
owe
ms Foster trust in AI through responsible government leadership in adoption, improving services, and leading by example.
r
Gov Use AI in government services
ern Enhance service quality, speed, and efficiency
Affir
men
mt
the
with
Use
AI Utilize sector-specific regulators with expertise in AI to engage with its uses effectively in different domains.
of a
Sect Maintain expertise in specialized agencies
oral Improve interagency coordination and access to AI resources
Reg
ulat
ory
Take
Stru
an Adopt an evolving policy framework to keep pace with AI’s rapid advancement, remaining flexible and vigilant over time.
ctur
Incr Stay adaptive as AI evolves
emee
Build a durable, evolving policy framework
ntal
App
roac
Keep
Humh Focus on human impact, ensuring the workforce is equipped for changes brought about by AI and prioritizing human
ans freedom and well-being.
at
the Address labor market effects
Cent Attract, train, and retain talent in AI
er of
AI
Polic
y

Key Findings Recommendations

 Government use of AI
 The federal government should utilize  Take an information and systems-level
core principles and avoid conflicting with approach to the use of AI in the federal
existing laws. government.
 The federal government should be wary  Support flexible governance.
of algorithm-informed decision-making.  Reduce administrative burden and
 The federal government should provide bureaucracy using AI.
notification of AI’s role in governmental  Require that agencies provide
functions. notification of AI’s role in governmental
 Agencies should pay attention to the functions.
foundations of AI systems.
 Facilitate and adopt AI standards for
 Roles and associated AI knowledge and
federal government use.
skills are unclear and highly varied across
 Support NIST in developing guidelines
the federal workforce.
for federal AI systems.
 Skills-based hiring is critical for filling
 Improve cybersecurity of federal
the demand for AI talent in the federal
 workforce. systems, including federal AI systems.
 Encourage data governance strategies
that support AI development.
 Congress and the government must
understand the federal government’s AI
workforce needs.
 Support different pathways into federal
service for AI talent.

Federal Preemption of State Law

 Federal preemption of state law on AI  Study applicable AI regulations across
issues is complex. sectors.
 Federal preemption has benefits and
drawbacks.
 Preemption can allow state action
subject to floors or ceilings.
 Preemption can be multifaceted.
 Definitions must be fit for purpose.

Data Privacy
 AI has the potential to exacerbate  Explore mechanisms to promote access
privacy harms. to data in privacy-enhanced ways.
 Americans have limited recourse for  Ensure privacy laws are generally
many privacy harms. applicable and technology-neutral.
 Federal privacy laws could potentially
augment state laws.

National Security
 AI is a critical component of national  Focus congressional oversight on AI
security. activities for national security.
 U.S. adversaries are adopting and  Support expanded AI training at DOD.
militarizing AI.  Continue oversight of autonomous
 National security requires advanced weapons policies.
cloud access and AI.  Support international cooperation on AI
 National security requires AI for used in military contexts.
contested environments.
 AI can vastly improve DOD business
processes.

Research, Development, & Standards

 Federal investments in fundamental  Continually monitor and evaluate the
research have enabled the current AI impact of AI on different industries and
opportunity. the nation.
 Continued AI research and evaluation  Support fundamental R&D for continued
will promote AI advancement. leadership in AI innovation.
 Progress in AI R&D is closely linked to  Increase technology transfer from
access to AI resources. university R&D to market.
 A closed AI research ecosystem could  Bipartisan House Task Force on
limit U.S. competitiveness in AI. Artificial Intelligence
 University AI R&D is necessary but  Promote public-private partnerships for
must be paired with vibrant technology AI R&D.
transfer activities.  Promote research and standardization
 Advancing the science around AI surrounding the evaluation and testing of
evaluation will help advance adoption. AI.
 The U.S. is a global leader in standard  Promote the development of
setting but faces competitors. infrastructure and data to enable AI
 research.
 Continue engagement in international
standards development.
 Uphold the U.S. approach to setting
standards.
 Align national AI strategy with broader
U.S. technology strategy.
 Explore how to accelerate scientific
discovery across disciplines with AI.
 Support AI R&D by small businesses.
 Encourage international collaboration
with like-minded allies and partners on
R&D.
Civil Rights & Civil Liberties
 Improper use of AI can violate laws and  Have humans in the loop to actively
deprive Americans of our most identify and remedy potential flaws when
important AI is used in highly consequential
rights. decision-making.
 Understanding the possible flaws and  Agencies must understand and protect
shortcomings of AI models can mitigate against using AI in discriminatory
potentially harmful uses of AI. decision-making.
 Empower sectoral regulators with the
tools and expertise to address AI-related
risks in their domains.
 Explore transparency for users affected
by decisions made using AI.
 Support standards and technical
evaluations to mitigate flawed decision-
making involving AI systems.

Education & Workforce

 AI is increasingly used in the workplace  Invest in K–12 STEM and AI Education
by both employers and employees. and Broaden Participation.
 Fostering domestic AI talent and  Bolster U.S. AI skills by providing
continued U.S. leadership will require needed AI resources.
significant improvements in basic  Develop a full understanding of the AI
STEM education and training. workforce in the United States.
 U.S. AI leadership would be  Facilitate public-private partnerships to
strengthened by utilizing a more skilled bolster the AI workforce.
technical workforce.  Develop regional expertise when
 AI adoption in America requires AI supporting government-university-
literacy. industry partnerships.
 K–12 educators need resources to  Broaden pathways to the AI workforce
promote AI literacy. for all Americans.
 Support the standardization of work
roles, job categories, tasks, skill sets, and
competencies for AI-related jobs.
 Evaluate existing workforce development
programs.
 Promote AI literacy across the U.S.
 Empower U.S. educators with AI training
and resources.
 Support NSF curricula development.
 Monitor the interaction of labor laws and
worker protections with AI adoption.
Intellectual Property
 It is unclear whether legislative action is  Clarify IP laws, regulations, and agency
necessary in some cases, and a number activity.
of IP issues are currently in the courts.  Appropriately counter the growing harm
 Generative AI poses a unique challenge of AI-created deepfakes.
to the creative community.
 It is often difficult for creators to know
if their copyrighted works are being
used by AI developers.
 The global IP policy landscape presents
challenges and opportunities to both
developers and creators.
 While some use cases are legitimate and
protected forms of expression, the
proliferation of deepfakes and harmful
digital replicas is a significant and
ongoing challenge.

Content Authenticity
 Synthetic content has many beneficial  Support a risk-based, multipronged
uses, but if used improperly it can create approach to content authenticity.
harms and undermine confidence in  Support technical solutions to content
information integrity. authenticity.
 There is currently no single, optimal  Address demonstrable harms, not
technical solution to content speculative harms of synthetic content.
authentication.  Identify the responsibilities of AI
 Technical literacy would help with the developers, content producers, and
content authenticity challenges but content distributors when it comes to
would not be sufficient. synthetic content.
 Digital identity technology allows a  Examine existing laws related to harmful
person online to verify who they are and synthetic content.
reduces fraud.  Ensure victims have the necessary tools.

Open & Closed Systems

 Open AI models encourage innovation  Encourage innovation and competition in
and competition. the development of AI models.
 There is currently limited evidence that  Focus on demonstrable harms and
open models should be restricted. physical threats.
 Evaluate chemical, biological,
radiological, or nuclear (CBRN) threats
in light of AI capabilities.
 Continue to monitor the risks from open-
source models.

Energy Usage & Data Centers

 AI is critical to U.S. economic interests  Support and increase federal investments
and national security and maintaining a in scientific research that enables
sufficiently robust power grid is a innovations in AI hardware, algorithmic
necessity. efficiency, energy technology
 The growing demands of AI are development, and energy infrastructure.
creating challenges for the grid.  Strengthen efforts to track and project AI
 Continued U.S. innovation in AI data center power usage.
requires innovations in the energy  Create new standards, metrics, and a
sector. taxonomy of definitions for
 Planning properly now for new power communicating relevant energy use and
generation and transmission is critical efficiency metrics.
for AI innovation and adoption.  Ensure that AI and the energy grid are a
 AI tools will play a role in innovation part of broader discussions about grid
and modernization in the energy sector. modernization and security.
  Ensure that the costs of new
infrastructure are borne primarily by
those customers who receive the
associated benefits.
 Promote broader adoption of AI to
enhance energy infrastructure, energy
production, and energy efficiency.

Small Business
 Small businesses can lack a full  Support small business AI literacy.
understanding of how best to adopt AI.  Provide resources for small business AI
 Small businesses can lack sufficient adoption.
access to capital and AI resources.  Investigate the resource challenges of
 Small businesses face excessive small businesses adopting AI.
challenges in meeting AI regulatory  Investigate the resource challenges of
compliance. small AI businesses.
 Ease compliance burdens for small
businesses.

Agriculture
 AI-driven precision agriculture could  Assess existing programs to identify
enhance farm productivity and natural opportunities for advancing AI in
resource management. precision agriculture.
 Increased AI integration could enable  Pursue further AI research and
mechanization and automation development to enhance efficiency in
technologies and enhance efficiency specialty crops.
within the specialty crop industry.  Continue to explore how research and
 Lack of reliable network connectivity in innovation in AI technology could aid
rural and farming communities impedes land managers in improving forest health
AI adoption in the agricultural sector. through better planning and strategies.
 AI is already a powerful tool in  Direct USDA to better utilize AI in
addressing and combating the wildfire program delivery.
and forest health crises.  Continue to review the application of the
 Greater adoption of AI at USDA could CFTC’s principles-based framework to
enhance delivery of numerous ensure it captures unique risks posed by
agriculture programs and reduce costs AI in financial markets.
for farmers and others.
 The CFTC’s principles-based approach
allows for flexibility in addressing new
technologies.

Health Care
 AI's use in healthcare can potentially  Encourage the practices needed to ensure
reduce administrative burdens and AI in healthcare is safe, transparent, and
speed up drug development and clinical effective.
diagnosis.  Maintain robust support for healthcare
 The lack of ubiquitous, uniform research related to AI.
standards for medical data and  Create incentives and guidance to
algorithms impedes system encourage risk management of AI
interoperability and data sharing. technologies in healthcare across various
deployment conditions to support AI
adoption and improve privacy, enhance
security, and prevent disparate health
outcomes.
 Support the development of standards for
liability related to AI issues.
 Support appropriate payment
mechanisms without stifling innovation.

Financial Services
 AI presents an opportunity to transform  Foster an environment where financial
the financial services sector. services firms can responsibly adopt the
 Data quality and data security are benefits of AI technology.
paramount in financial service AI  Encourage and resource regulators to
models. increase their expertise with AI.
 AI can expand access to financial  Maintain consumer and investor
products and services. protections in the use of AI in the
 AI technologies are already deployed financial
across the financial services sector. services and housing sectors.
 Some regulators use AI to identify non-  Consider the merits of regulatory
compliance with regulations. “sandboxes” that could allow regulators
 Small financial services firms can be at to experiment with AI applications.
a disadvantage in AI adoption.  Support a principles-based regulatory
approach that can accommodate rapid
technological changes.
 Ensure that regulations do not impede
small firms from adopting AI tools.