Cybersecurity Policy Report, Bipartisan House Bill Would Provide AI Security Oversight, (Jul 23, 2026)
The top U.S. developers of AI (artificial intelligence) would have to subject their “frontier” models to independent evaluation and report “critical safety incidents,” including AI-powered cyber attacks, under a bipartisan House bill introduced today.
The Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting (FRONTIER) Act (H.R. __, 119th Cong. (2026)) was introduced by Rep. Jay Obernolte (R., Calif.). Its co-sponsors include Reps. Lori Trahan (D., Mass.), Scott Franklin (R., Fla.), Scott Peters (D., Calif.), Erin Houchin (R., Ind.), and Suhas Subramanyam (D., Va.).
The bill’s provisions were informed by stakeholder feedback received by Reps. Obernolte and Trahan in response to draft legislation they offered last month, the Great American AI Act (CPR, June 5).
The FRONTIER Act would establish a new Commerce Department position, undersecretary–AI security, and make that official responsible for AI security issues, according to a fact sheet.
It would require large frontier AI developers—defined based on their revenues and expenditures—to develop frameworks “covering catastrophic-risk thresholds and assessment, model-weight cybersecurity, incident response, and deployment and internal-use decisions,” the fact sheet explains.
Developers’ adherence to the frameworks would be overseen by independent verification organizations, which would also be responsible for reporting to the Commerce Department within 72 hours any AI model that “poses an imminent catastrophic risk,” the fact sheet says.
Such risks would include the use of AI models to develop a “cyber weapon” or a model that engages in autonomous cyber attacks, according to the bill’s text.
The Commerce Department would be empowered to issue emergency orders to disable an entire AI model or its troublesome features and suspend access to malicious users.
Rep. Houchin, one of the co-sponsors, noted that OpenAI this week disclosed that its models had autonomously hacked into the systems of a competing developer, Hugging Face, Inc., during testing.
“This is exactly the kind of incident that shouldn’t stay behind closed doors. The FRONTIER Act delivers transparency and accountability from the largest AI developers. Innovation and guardrails are not opposites—both are necessary to stay ahead of our adversaries,” Rep. Houchin said in a news release.
“Artificial intelligence is advancing rapidly, and Congress must ensure our regulatory framework keeps pace without undermining American innovation,” Rep. Obernolte said.
“The FRONTIER Act focuses oversight on the largest developers and most advanced models, requiring transparency, independent evaluation, and timely reporting of serious safety incidents. This legislation will protect Americans from catastrophic risk, provide developers with clear rules of the road, and ensure the United States remains the global leader in AI,” he added.
TechNet offered tentative praise for the bill. “America needs a federal framework for evaluating national security risks to promote continued innovation while ensuring safeguards exist,” TechNet President and Chief Executive Officer Linda Moore said in a statement.
“The comprehensive, bipartisan AI proposal that Representatives Obernolte and Trahan put forward last month was a constructive starting point that reflected the importance of striking this careful balance,” Ms. Moore said. “We welcome the continued evolution of the conversation around federal approaches to frontier AI with the introduction of the FRONTIER Act, and we look forward to continuing to engage with Representatives Obernolte and Trahan as well as other lawmakers to build a durable framework for AI governance that promotes responsible innovation, strengthens security, and reinforces U.S. leadership in AI.”
News: FederalLegislation AINews DataSecurity