Bill to require development of federal AI security playbook

FedScoop reports that the Advanced AI Security Readiness Act sponsored by Sens. Mark Kelly, D-Ariz., and Todd Young, R-Ind., would mandate the National Security Agency to develop a security playbook to protect U.S. artificial intelligence systems from foreign threats.

The bipartisan bill directs the NSA's AI Security Center to outline vulnerabilities in advanced AI technologies and their supply chains, while also laying out strategies to defend critical computing environments. The playbook is expected to address risks such as model-weight theft, insider threats, and cyberespionage attempts aimed at weakening the confidentiality or reliability of U.S. AI systems.

According to the legislation, the NSA must work closely with major AI developers, national laboratories, technical experts, and federal agencies, including the Department of Defense, the National Institute of Standards and Technology's Center for AI Standards and Innovation, the Commerce Department's Bureau of Industry and Security, and the Department of Homeland Security.

The senators said the guidance is necessary as adversaries race to steal or exploit American innovation. The effort follows recent NSA publications on securing AI systems, alongside heightened public scrutiny of the agency's own AI use.