After a year of anticipation, Senate Majority Leader Chuck Schumer (D-NY) released the Bipartisan AI Working Group’s long-awaited roadmap for AI policy titled “Driving U.S. Innovation in Artificial Intelligence” last week. The Working Group was established in May 2023 to serve as the driving force behind Congress’s response to AI. The group has engaged in hundreds of stakeholder meetings and hosted nine AI Insight Forums, featuring conversations with tech CEOs to understand their perspective. The group’s work has focused on issues including workforce displacement, national security, intellectual property, election disinformation, and domestic innovation. The group has also attempted to balance addressing the dangers posed by AI with the need to ensure that the United States remains strategically competitive against China and other potentially hostile, where the AI market is expected to exceed $26 billion by 2026.
Within this report, the working group sought to tackle bipartisan opportunities for comprehensive AI regulation that warrant enhanced committee consideration. Collectively, the group proposes $32 billion in annual non-defense spending dedicated toward AI R&D initiatives that seek to position the United States as a global leader in AI. The proposal splits the $32 billion of funding across:
– Department of Commerce – Department of Energy – National Science Foundation – National Institute of Health – National Institute for Standards and Technology – National Aeronautics and Space Administration
While these recommendations are voluntary, the roadmap is meant to provide committees of jurisdiction with direction in how they craft future legislation. Schumer envisions this report as a blueprint for how AI will be regulated in the coming years, focusing on areas where strong bipartisan consensus exists.
– Supporting domestic innovation in AI R&D
– Identifying and addressing national security risks and opportunities through increased deployment of trustworthy AI systems within the DoD
– Providing training and upskilling resources to the private sector workforce to ensure participation in the AI-enabled economy
– Recruiting high-skilled tech talent to the federal workforce
– Imposing testing, evaluation, and reporting standards on high-impact AI systems
– Addressing the usage of AI systems to produce harmful content, perpetuate fraud and deception, and endanger children
– Protecting private consumer information via a comprehensive data privacy law
– Advancing watermarking and digital content provenance to prevent the spread of AI-generated disinformation ahead of the 2024 election
– Imposing disclosure requirements and limitations on AI training data sets to protect the intellectual property of content creators
– Monitoring the deployment of adversarial AI and considering whether export controls should be imposed on “powerful” AI systems
– Expanding collaboration with international allies to encourage open markets and the free flow of information spearheaded by American companies to combat China and Russia’s efforts to restrict the digital economy
Several existing AI bills were referenced in the report. These bills to watch include:
CREATE AI Act (2714 / H.R. 5077) Establishes the National Artificial Intelligence Research Resource (NAIRR) to provide AI researchers with expanded access to complex data, tools, and resources to develop safe and trustworthy AI Future of AI Innovation Act (4178) Establishes AI standards, metrics, and evaluation tools to support AI R&D via the NIST AI Safety Institute, publicly available AI datasets, international partnerships, grand challenge prize competitions, and AI testbeds within National Laboratories AI Grand Challenges Act (4236) Establishes an AI Grand Challenges Program within the National Science Foundation (NSF) to provide researchers and entrepreneurs with a minimum $1 million prize for projects that utilize AI to solve problems that benefit the public good Small Business Technological Advancement Act (2330) Expands the Small Business Administration’s (SBA) 7(a) loan program to provide small businesses with access to technology that improves daily operations Workforce DATA Act (2138 / H.R. 6518) Requires the Bureau of Labor Statistics (BLS) to analyze the impact of automation on the workforce to support workforce development strategies Artificial Intelligence Advancement Act of 2023 (3050) – Section 3 Requires an AI regulatory gap analysis in the financial sector AI Transparency in Elections Act of 2024 (3875) Requires disclaimers on political advertisements with AI-generated content Preparing Election Administrators for AI Act (3897 / Introduced in House) Directs the Election Assistance Commission, in collaboration with the National Institute of Standards and Technology (NIST), to develop guidelines that will assist election administrators in addressing AI’s impact on elections Protect Elections from Deceptive AI Act (2770) Bans the usage of AI to generate false content depicting federal candidates in political advertisements to influence the outcome of a federal election
At a press conference announcing the new report, Schumer stated that he plans to meet with House Speaker Mike Johnson (R-LA) to discuss opportunities for collaboration. Schumer has also made it clear that he does not envision a comprehensive AI legislative package. Instead, he anticipates that many of the policy areas highlighted within the report will be advanced via committee consideration of individual AI bills.
While Schumer has acknowledged the limited runway for AI legislation this year, his top priority is addressing the use of AI in elections. Senators are moving full speed ahead with this effort. Hours after the report’s release, Senate Rules Committee Chair Amy Klobuchar (D-MN) announced the successful markup of three AI election bills that would:
– Ban deepfakes falsely depicting federal candidates in political advertisements – Require disclaimers on political advertisements that utilize AI-generated content – Institute guidelines that prepare election workers to address the impact of AI in election administration, disinformation, and cybersecurity
Schumer remains committed to passing election-focused AI legislation before the end of the 118th Congress. In the coming weeks, both Schumer and the broader Working Group, including Senators Martin Heinrich (D-NM), Todd Young (R-IN), and Mike Rounds (R-SD), plan to meet with individual committees to advocate for legislative activity aligned with the report recommendations.
A majority of the recommendations in this roadmap likely will not be materialized into bills that can clear both chambers before the end of the year. With a narrowly divided House, a lack of consensus in the Senate, and only three months left in the legislative calendar of the 118th Congress before the 2024 elections, it is unlikely that most of the legislative recommendations outlined above will be passed into law until the 119th Congress.
If any AI legislation were to move this year, it would likely be related to addressing AI-generated disinformation ahead of the 2024 election. While the three bills passed during last week’s markup will similarly face significant hurdles to passage, Schumer has made it clear that addressing AI usage in elections is his top priority for the 118th Congress. In the Senate, Republicans have expressed concerns with how one bill, the Protect Elections From Deceptive AI Act (S.2770), may conflict with existing state laws to combat deepfakes and the bill’s potential infringement on Americans’ First Amendment rights. While Senate passage does not guarantee success, as the House may drag its feet on any movement ahead of the election, Schumer’s upcoming conversation with Speaker Johnson will provide more clarity on the path forward.
In the next few months, lawmakers will prioritize passing the remaining major legislative packages for the year, including the FY 2025 appropriations bills, the Farm Bill, and the National Defense Authorization Act (NDAA). Schumer has indicated his interest in attaching AI legislation to the NDAA ahead of the Senate markup later this month, which may provide Klobuchar’s election bills with a legislative vehicle for passage this Congress.
Outside of Congress, expect the Biden administration to similarly push for executive actions to prevent the spread of deepfakes. While the president’s authority to address AI-generated disinformation through executive orders is limited, Biden’s most tangible path forward would be to issue guidance to election administrators akin to the recommendations within Klobuchar’s Preparing Election Administrators for AI Act. The Defense Advanced Research Project Agency (DARPA) is also expected to play a key role in these enforcement efforts.
The new roadmap will serve as a guiding document for Members as they continue to draft AI legislation. While the recommendations are non-binding, the broad strokes painted within this roadmap will provide committees with the flexibility to collect information and develop legislation that builds upon each of the report’s policy prescriptions. In his remarks to the Center for Strategic and International Studies (CSIS) last year, Schumer called on Congress to, “come up with a plan that encourages – not stifles – innovation in this new world of AI.” Schumer recognizes that, while AI development will proceed at an exponential pace, Congress should not rush to pass overbearing laws that could hamper U.S. innovation and national security.