Highlight any text to annotate

# Summary of S. 5061: Secure A.I. Development Act of 2026 This legislation establishes a comprehensive federal framework for monitoring and securing artificial intelligence systems before they are released to the public. The bill creates an Artificial Intelligence Risk Board within the National Institute of Standards and Technology that will develop technical evaluations and best practices for frontier AI models—those with advanced capabilities that could pose risks to national security, the economy, or public safety. Companies must provide the National Security Agency with access to frontier AI models at least 21 days before public release, and must register these models in a government registry. The bill also establishes a voluntary database for reporting AI security and safety incidents, allowing companies, researchers, and civil society groups to anonymously report problems while maintaining confidentiality protections. Additionally, it directs federal agencies to update their cybersecurity vulnerability tracking systems to accommodate AI-related vulnerabilities, which are expected to increase significantly. The legislation includes a three-year pilot program allowing the National Security Agency to share classified threat intelligence with AI developers working on federal contracts, helping them protect against foreign espionage and sabotage. Companies that violate the requirement to provide pre-release access to frontier models face fines of at least $100,000 per day, though they have seven days to comply after receiving notice. The bill requires implementation within 90 to 180 days for most provisions.

Add a comment

Sign in to comment on this bill.

Annotate the text

Highlight any passage on the Summary or Full Text tab to attach a note. Annotations appear on the Annotations tab.