Generative AI Terrorism Risk Assessment Act

Annual reports on generative AI terrorism

This bill would require the Secretary of Homeland Security, with the Director of National Intelligence, to give Congress a written assessment about generative AI terrorism threats within one year of enactment and once a year for five years. Each assessment would analyze the prior calendar year for cases where foreign terrorist organizations or individuals used or tried to use generative AI to spread violent messages, recruit or radicalize people, or help develop or deploy chemical, biological, radiological, or nuclear weapons. Each report would include recommendations, be coordinated to protect privacy and civil liberties, be unclassified with an optional classified annex, and the unclassified portion would be posted on DHS's public website. DHS would brief the appropriate congressional committees within 30 days of each submission and would review, use, and share information from State and major urban area fusion centers, the National Network of Fusion Centers, ODNI, the FBI, and other federal agencies. For purposes of the bill, "foreign terrorist organization" would mean groups listed under immigration law, and "generative artificial intelligence" would mean AI models that create synthetic images, video, audio, text, and other digital content.