Sponsor
Mark R. Warner
Senator · D-VA
Senator · D-VA
The bill strengthens protections, technical standards, and civil remedies to reduce abusive synthetic content and protect victims, but it raises liability, compliance costs, and risks to small vendors and lawful research and expression.
Children and survivors: federal agencies must block AI models that generate child sexual abuse material or non‑consensual intimate images, increasing removal and protection.
Individuals depicted in abusive synthetic content: a new civil remedy (including statutory damages and injunctions) lets victims seek redress and removal.
Tech vendors and buyers: NIST will create performance benchmarks and a voluntary vendor test program within 180 days to reduce abusive and obscene synthetic content, giving clearer standards for compliance and procurement.
Vendors and developers: exposure to sizable statutory damages (e.g., $50,000) and liability increases legal risk and potential financial harm for companies and creators.
Federal agencies and contractors: procurement and compliance burdens may increase, potentially forcing removal or replacement of AI tools within tight timelines and raising costs.
Small vendors and startups: difficulty meeting evolving NIST benchmarks and voluntary testing could reduce competition and slow innovation in the AI market.
Based on analysis of 2 sections of legislative text.
Bars federal use or procurement of generative AI models that fail NIST's Generative AI Profile or have been judicially found to produce child pornography; tasks NIST with benchmarks and voluntary testing.
Official title: Prohibit the use of certain artificial intelligence models across the Federal Government, and for other purposes.
Introduced July 21, 2026 by Mark R. Warner · Last progress July 21, 2026
Prohibits federal agencies from procuring or using generative AI models that fail to meet NIST’s Generative AI Profile or have been judicially found to produce child pornography, and requires removal of such models from agency and contractor systems within 180 days. Directs NIST to develop performance benchmarks and a voluntary vendor testing program to reduce generation of non‑consensual intimate images and synthetic child sexual abuse material, with initial actions starting within 90–180 days of enactment. The bill defines covered terms, requires continuous agency evaluation and removal of disallowed models, and centers federal procurement and use rules on alignment with the NIST AI Risk Management Framework (Generative AI Profile) and related technical controls aimed at preventing abusive and illegal content generation.