Sponsors (2)
The bill strengthens protections, standards, and legal remedies against abusive synthetic content—better protecting children and victims and pushing vendors toward safer AI—but does so at the cost of added compliance and liability burdens, potential market consolidation, trade/enforcement complexity, and risks of chilling legitimate research and expression.
Children and survivors of non-consensual intimate-image and sexual-abuse synthetic content are better protected because federal agencies must block AI models that generate such material.
Developers and vendors are pushed toward stronger safety and accountability through NIST-created performance benchmarks and a rapid (180-day) voluntary vendor test program, encouraging technical controls and standardized testing for abusive or obscene synthetic content.
Individuals depicted in abusive synthetic content gain a civil remedy (including statutory damages and injunctions), giving victims a clearer path to removal and compensation.
Researchers, artists, and students may face a chilling effect on legitimate research and expressive uses of generative models because broad prohibitions and civil claims risk suppressing lawful speech and investigation.
Vendors and developers face substantial new legal risk and potential statutory damages (e.g., $50,000), increasing liability exposure, compliance costs, and insurance/legal burdens for businesses offering generative models.
Small vendors and startups may struggle to meet evolving NIST benchmarks and testing expectations, risking reduced competition and innovation as market concentration favors larger firms that can absorb compliance costs.
Based on analysis of 2 sections of legislative text.
Bars federal procurement/use of generative AI models that fail NIST benchmarks for preventing non-consensual intimate imagery or are judicially found to have generated child pornography, and directs NIST to create benchmarks and vendor tests.
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 standards for preventing non-consensual intimate imagery and synthetic child sexual abuse material or that have been judicially found to generate child pornography. Directs NIST to develop performance benchmarks and a voluntary vendor test program to assess models for risks of producing obscene, abusive, or exploitative content. Requires agencies and their contractors to remove covered applications from government systems within 180 days, establishes definitions and technical-control requirements, and mandates ongoing evaluation by agencies and GSA to prevent use of models that generate non-consensual intimate images or synthetic child sexual abuse material.