Sponsors (10)
The bill strengthens government and industry tools to prevent and mitigate AI‑related security harms and to coordinate on urgent threats, but does so at the cost of higher compliance and litigation burdens, reduced transparency, potential competition harm, and broader enforcement authority that could constrain developers and researchers.
General public and consumers: the Attorney General can seek court injunctions to stop non‑Federal actors whose actions are likely to increase covered AI security risks, enabling faster legal remedies to reduce public harms.
Operators of critical infrastructure (utilities, hospitals, energy companies): clearer statutory definitions help target AI risks, improving risk assessment and mitigation planning for systems that affect public health and safety.
AI developers, deployers, and model owners: the bill clarifies key legal terms (e.g., covered AI security risks, unauthorized access) so organizations can better prioritize security actions, support incident response, and pursue legal remedies against intrusions and model theft.
AI developers, model owners, and companies: new compliance requirements and greater litigation and injunction risk (when asserting AI security purposes) will increase legal and operational costs and may chill innovation or raise consumer prices.
Competition and consumers: the safe harbor that allows firms to coordinate delays or limits on deployment creates a pathway that could reduce market competition, slow product rollouts, and enable tacit or subtle collusion despite prohibitions.
Researchers, developers, and civil liberties advocates: broad risk language and expanded DOJ authority over non‑Federal actors could enable wide enforcement discretion, export or usage restrictions, and potential overreach into legitimate research and development activities.
Based on analysis of 4 sections of legislative text.
Creates a narrow antitrust safe harbor for exclusive, good‑faith information sharing or temporary coordination to reduce defined AI security risks, with DOJ notice and FOIA protections.
Official title: To establish the applicability of antitrust laws to the sharing of artificial intelligence frontier model risks, and for other purposes.
Introduced July 23, 2026 by Robert E. Latta · Last progress July 23, 2026
Creates a narrow antitrust safe harbor allowing non‑Federal entities to share information or temporarily coordinate delays or limits on AI release, deployment, development, training, testing, or evaluation when done exclusively to reduce defined AI security risks. The exemption requires good faith, written notice to the Antitrust Division for coordination, internal controls, and is an affirmative defense subject to limits; classic anticompetitive conduct remains unlawful and the Attorney General may seek injunctions against misuse. Defines key terms related to artificial intelligence security risks (including theft by covered nations, facilitation of WMD or cyberweapons, threats to critical infrastructure, loss of governance, autonomous self‑improvement risks, and vulnerabilities enabling unauthorized access) and narrows what counts as authorized information exchange or coordination, with FOIA protection for required notices to the Antitrust Division.