Recently, a bipartisan group of U.S. senators introduced new legislation to address transparency and accountability for artificial intelligence (AI) systems, including those deployed for certain “critical impact” use cases. While many other targeted, bipartisan AI bills have been introduced in both chambers of Congress, this bill appears to be one of the first to propose specific legislative text for broadly regulating AI testing and use across industries.
The Artificial Intelligence Research, Innovation, and Accountability Act—led by Senate Commerce Committee members John Thune (R-SD), Amy Klobuchar (D-MN), Roger Wicker (R-MS), John Hickenlooper (D-CO), Shelly Moore Capito (R-WV), and Ben Ray Luján (D-NM)—would establish new testing standards for high-impact AI systems; mandate reporting from AI companies on the testing, training, use, and benefits of high-impact AI systems; and formalize disclosure requirements for AI-generated content. The bill would also create additional safeguards for certain “critical-impact” AI systems, including those that involve the collection and processing of biometric data, relate to critical infrastructure, or have criminal-justice applications.
The Thune-Klobuchar bill would also direct the National Institute of Standards and Technology to facilitate standards for capturing and disclosing the chain of development (known as “provenance”) of digital content, which would allow users to understand whether AI was involved in producing particular content and assess content authenticity.
This new legislation was introduced as Congress and the Administration continue their robust focus on AI policymaking. Majority Leader Chuck Schumer (D-NY) and a bipartisan group of colleagues announced their “SAFE Innovation Framework” in June, and have continued to host AI Insight Forums with industry experts and stakeholders. Senators Richard Blumenthal (D-CT) and Josh Hawley (R-MO) announced their privacy framework for AI regulation in September.
Dozens of targeted bipartisan AI bills are also pending in Congress, including more than half a dozen that have progressed through their committees of jurisdiction. And, in October, President Biden released his AI executive order outlining a comprehensive strategy to support the development and deployment of safe and secure AI.