Senate AI Working Group Releases Long-Anticipated Roadmap for AI Policy

Key Points

• On May 15, 2024, the Senate AI Working Group—Senate Majority Leader Chuck Schumer (D-NY) and Sens. Mike Rounds (R-SD), Todd Young (R-IN), and Martin Heinrich (D-NM)—issued their long-anticipated Roadmap for Artificial Intelligence (AI) Policy in the United States Senate.
• The Roadmap contains recommendations for Senate committees of jurisdiction on sector-specific AI policy issues as they craft legislation, with the goal of reaching bipartisan consensus on key issues covered in the previously convened Senate AI Insight Forums.
• Moving ahead, the Working Group encourages Senate committees to closely collaborate with one another on legislation, including by reaching agreement on key definitions, and the lawmakers also urge the Biden Administration to update Congress on executive-led initiatives, including Memorandums of Understanding (MOUs) with other countries.

Background

On May 15, 2024, the Senate AI Working Group—Senate Majority Leader Chuck Schumer (D-NY) and Sens. Mike Rounds (R-SD), Todd Young (R-IN), and Martin Heinrich (D-NM)—issued their long-anticipated Roadmap for Artificial Intelligence (AI) Policy in the United States Senate (“Roadmap”) with recommendations for Senate committees of jurisdiction on sector-specific AI policy issues as they craft legislation, with the goal of reaching bipartisan consensus on key issues covered in the previously-convened Senate AI Insight Forums.

The 118th Congress has seen an unprecedented amount of attention on AI. Committees in the House and Senate have held more than 50 hearings this year to examine AI’s impact on a broad range of issues including national security, cybersecurity, elections, energy, healthcare, financial services, consumer protection and intellectual property. Through 17 months this Congress, lawmakers have introduced more than 80 individual pieces of legislation related to AI. By comparison, the previous Congress saw only 21 pieces of standalone AI legislation in two years.

The broad range of topics that need to be addressed and the myriad potential impacts of AI have led to the formation of several informal entities within Congress seeking to bring coherence and focus to an otherwise all-encompassing legislative process. In April of last year, Leader Schumer announced one such effort to develop legislation to regulate AI. Throughout last fall, Leader Schumer, alongside Sens. Heinrich, Young and Rounds, convened a series of “AI Insight Forums” to seek input from industry, academia and civil society.

The nine Insight Forums focused on (1) the overarching issues that the government needs to address as it attempts to regulate; (2) American leadership on AI; (3) AI workforce considerations; (4) AI’s impact on high-risk applications such as healthcare and finance; (5) democracy and elections; (6) privacy and liability; (7) transparency in AI and the intellectual property (IP) protections that creators should be afforded; (8) “doomsday scenarios”, including societal risks associated with the development and deployment of AI; and (9) national security considerations.

The Roadmap maintains an overarching focus on innovation and federal investment to promote sustainable technological development. While it is not intended to be exhaustive, the Roadmap will serve as a high-level blueprint for Senate committees of jurisdiction as they continue to develop and advance bipartisan legislative proposals.

Roadmap Summary

Supporting U.S. Innovation in AI

• Appropriations:
  • This section maintains an emphasis on the need for increased federal investment in AI, echoing Leader Schumer’s previously stated goal of securing at least $32 billion per year for non-defense Al innovation, as proposed by the National Security Commission on Artificial Intelligence (NSCAI)’s final report. To begin the process of reaching this goal, the Roadmap encourages the Senate Appropriations Committee to work with the relevant committees to develop emergency appropriations language.
  • The Roadmap highlights a number of priority areas for funding, including a cross-government Al research and development (R&D) effort; the outstanding CHIPS and Science Act (P.L. 117-167) accounts; semiconductor R&D specific to the design and manufacturing of future generations of high-end Al chips; funding for Al efforts at the National Institute of Standards and Technology (NIST), including Al testing and evaluation infrastructure and the U.S. Al Safety Institute; a NIST and US. Department of Energy (DOE) testbed; and updates to the federal government’s IT infrastructure.
• Endorsement of Existing Legislation:
  • CREATE Al Act (S. 2714): Introduced by Sen. Heinrich, the bill would authorize the National Al Research Resource (NAIRR), which the Roadmap outlines the need to fund as part of the cross-government Al initiative. The Roadmap also calls for expanding programs such as the NAIRR and the National Al Research Institutes.
  • Small Business Technological Advancement Act (S. 2330): Introduced by Sen. Young, the bill would clarify that business software and cloud computing services are allowable expenses under the SBA's 7(a) loan program.
  • Section 202 of the Future of Al Innovation Act (S. 4178) and the Al Grand Challenges Act (S. 4236): The Working Group also supports “Al Grand Challenge” programs described in the two bills.
• Proposed Legislative Reforms:
  • Among the Roadmap’s additional recommendations to drive innovation are directives for the committees of jurisdiction to develop legislation to leverage public-private partnerships across the federal government; consider whether legislation is needed to support best practices to incentivize states and localities to invest in similar opportunities as those provided by the NAIRR; and work with the U.S. Department of Commerce (Commerce) and other agencies to increase access to tools, such as mock data sets, for Al companies to utilize for testing.
• Proposed Executive Actions:
  • The Roadmap calls for a report from the Comptroller General of the United States to identify any significant federal statutes and regulations that affect AI innovation, including the ability of companies to compete.

AI and the Workforce

• Endorsement of Existing Legislation:
  • Workforce Data for Analyzing and Tracking Automation Act (S. 2138): The bill, which was introduced by Senate Homeland Security Chair Gary Peters (D-MI), would authorize the Bureau of Labor Statistics (BLS) to record the effect of automation on the workforce and measure those trends over time, including job displacement, the number of new jobs created and the shifting in-demand skills. The bill would also establish a workforce development advisory board composed of key stakeholders to advise the U.S. Department of Labor (DOL) on public and private sector initiatives to promote workforce development improvements.
• Proposed Legislative Reforms:
  • The Roadmap’s additional recommendations to mitigate potential adverse impacts of Al on the workforce include ensuring that stakeholders are consulted as Al is developed and deployed by end users, or procured or used by federal agencies; crafting legislation related to training, retraining and upskilling the domestic workforce; developing a policy framework addressing the impact of Al on long-term future of work, and considering legislation to improve the U.S. immigration system for high-skilled STEM workers.
• Proposed Executive Actions:
  • The Roadmap calls on the federal government to, in its adoption of Al, improve government service delivery and modernize internal governance as well as upskilling of existing federal employees and exploration of opportunities to recruit and retain talent in Al through programs like the U.S. Digital Service.

High-Impact Uses of AI

• Endorsement of Existing Legislation:
  • Artificial Intelligence Advancement Act (S. 3050): The Roadmap supports Section 3 of the bill, which would direct a regulatory gap analysis in the financial sector.
• Proposed Legislative Reforms:
  • The Roadmap encourages relevant committees to develop legislation that ensures financial service providers are using accurate and representative data in their Al models.
  • The Roadmap calls on committees to advance standards for use of Al in critical infrastructure; examine the risks of the use of Al systems in the housing sector; address the Al-related concerns of professional content creators and publishers; develop legislation to address online child sexual abuse material (CSAM) and non-consensual distribution of intimate images; consider legislation to protect children from potential Al-powered harms online; and explore mechanisms, including through the use of public-private partnerships, to deter the use of Al to perpetrate fraud and deception.
  • The Roadmap contains a range of health policy priorities, including implementing appropriate guardrails and safety measures to protect patients; supporting the National Institutes of Health (NIH) in the development of Al technologies; providing for transparency in the use of Al in medical products and clinical support services, including the data used to train the Al models; and examining the Centers for Medicare & Medicaid Services’ (CMS) reimbursement mechanisms to promote innovation and ensure accountability.
• Proposed Executive Actions:
  • The Roadmap also outlines the need for the federal government to ensure appropriate testing and evaluation of Al systems in the federal procurement process and streamline the federal procurement process for Al systems and other software. It further calls on the Energy Information Administration (EIA) to include data center and supercomputing cluster energy use in their regular voluntary surveys.

Elections and Democracy

• Proposed Legislative Reforms:
  • The Roadmap encourages relevant committees and Al developers and deployers to advance effective watermarking and digital content provenance as it relates to Al-generated or Al-augmented election content.
• Proposed Executive Actions:
  • The Working Group acknowledges the U.S. Election Assistance Commission (EAC) for its work on the Al Toolkit for Election Officials, and the Cybersecurity and Infrastructure Security Agency (CISA) for its work on the Cybersecurity Toolkit and Resources to Protect Elections and encourages states to consider utilizing the tools EAC and CISA have developed.

Privacy and Liability

• Proposed Legislative Reforms:
  • The Roadmap outlines the need for a comprehensive federal privacy standard and encourages committees to explore policy mechanisms to reduce the prevalence of non-public personal information being stored in, or used by, Al systems. The Roadmap also suggests providing appropriate incentives for research and development of privacy-enhancing technologies.
  • The Roadmap also directs committees to consider whether there is a need for additional standards to hold Al developers and deployers accountable if their products or actions cause harm to consumers, or to hold end users accountable if their actions cause harm, as well as how to enforce any such liability standards.

Transparency, Explainability, IP and Copyright

• Proposed Legislative Reforms:
  • The Roadmap directs committees to consider developing legislation to establish public-facing transparency requirements for Al systems; evaluate whether there is a need for best practices for the level of automation for a given type of task; review how federal agencies are required to provide transparency to their employees about the deployment of new technology in the workplace; consider policy issues related to the data sets used by Al developers to train their models; review forthcoming reports from the executive branch related to establishing provenance of both synthetic and non-synthetic digital content; consider legislation that incentivizes software providers using generative Al and hardware products to provide content provenance information; and consider legislation requiring online platforms to maintain access to such content provenance information.
  • Regarding IP and copyright, the Working Group calls on committees to consider legislation protecting against the unauthorized use of one's name, image, likeness and voice; and review the results of reports from the U.S. Copyright Office and the U.S. Patent and Trademark Office (USPTO) on how Al impacts copyright and intellectual property law and take action if necessary.
  • In terms of explainability, the Roadmap calls on committees to consider legislation establishing a public awareness and education campaign for the benefits and risks of Al in the daily lives of Americans.

Safeguarding Against AI Risks

• Proposed Legislative Reforms:
  • The Roadmap directs committees to consider a resilient, risk-based regime that focuses on the capabilities of Al systems, protects proprietary information and allows for continued Al innovation in the U.S. In particular, the lawmakers outline the importance of development and standardization of risk testing and evaluation methodologies and mechanisms, including red-teaming, sandboxes and testbeds, commercial Al auditing standards, bug bounty programs, as well as physical and cybersecurity The Roadmap calls on committees to develop an analytical framework that specifies what circumstances would warrant a requirement of pre-deployment evaluation of Al models and develop legislation aimed at advancing R&D efforts that address the risks posed by Al system capabilities. It also calls on committees to examine the full spectrum of product release choices between closed and fully open-source models and explore whether there is a need for an Al-focused Information Sharing and Analysis Center (ISAC) to serve as an interface between commercial Al entities and the federal government regarding Al risks.

National Security

• Proposed Legislative Reforms:
  • The Roadmap calls on committees to develop legislation to bolster the use of Al in U.S. cyber capabilities; address artificial general intelligence (AGI); improve lateral and senior placement opportunities to expand the Al talent pathway into the military; develop a framework for determining when export controls should be placed on powerful Al systems; develop a framework for determining when an Al system, if acquired by an adversary, would be powerful enough that it would pose a grave risk to national security; and participate in international partnerships with like-minded allies and partners.
  • The Working Group also encouraged committees to assess whether aspects of the DOD’s policy on fully autonomous lethal weapon systems should be codified; explore leveraging advanced Al models to improve the risk mitigation of space debris; mitigate the rising energy demand of Al systems; ensure the Bureau of Industry and Security (BIS) proactively manages critical technologies; consider the recommendations of the National Security Commission on Emerging Biotechnology and the NSCAI; and work with the executive branch to support the free flow of information across borders.
• Proposed Executive Actions:
  • The Roadmap encourages the U.S. Department of Defense (DoD) and the U.S. Intelligence Community (IC) to develop career pathways and training programs for digital engineering, specifically in Al, as outlined in Section 230 of the fiscal year (FY) 2020 National Defense Authorization Act (NDAA; P.L. 116-92). It also calls on the Office of the Director of National Intelligence (ODNI), DoD and DOE to work with commercial Al developers to prevent large language models, and other frontier Al models, from inadvertently leaking or reconstructing sensitive or classified information.

Conclusion

Senate committees will now be tasked with exploring these issues and, where possible, developing bipartisan legislative solutions to address them. Leader Schumer has stated that the Senate will not wait to take up a single omnibus AI bill. Instead, the body will aim to take up and pass legislation on a rolling basis as opportunities arise. Areas to monitor include FY 2025 appropriations, the National Defense Authorization Act and bipartisan legislation that has previously been introduced in the Senate Commerce Committee.

Prior to the Roadmap release, committees continued to press ahead on AI. In particular, the Roadmap release coincided with the Senate Rules Committee’s advancement of three bills introduced by Chair Amy Klobuchar (D-MN) to address the impact of AI-generated deepfakes on elections: the Protect Elections from Deceptive AI Act (S. 2770), AI Transparency in Elections Act of 2024 (S. 3875), and the Preparing Election Administrators for AI Act (S. 3897).

The Akin cross-practice AI team continues to advise clients on navigating the evolving AI regulatory landscape and will closely track Congressional efforts to regulate AI, as well as implementation of the Biden-Harris Administration’s directives and the resulting opportunities for industry engagement, and keep clients apprised of key developments.