High-risk AI models would be tested by the government and registered before public release.
This bill would create new rules to ensure powerful artificial intelligence (AI) models are tested for safety and security before they are available to the public. It would establish a board to define high-risk AI and require companies to give the National Security Agency access to these models for testing. The bill would also create a public database to track AI security and safety incidents, helping to identify and address potential problems.
Today, there isn't a clear, mandatory system for testing powerful AI models for security risks before they are released to the public. There's also no central, public database for tracking AI safety and security incidents. After this bill, developers of high-risk AI models would need to allow the National Security Agency to test their systems and register them, and a new public database would collect information on AI incidents to help prevent future problems.
S 5061 · 119th Congress · July 21, 2026 · AI Summary by gemini-2.5-flash · 10/10
Sign in to see your representatives' phone numbers