OpenAI has publicly endorsed a bipartisan House bill that would establish the first federal framework for AI safety, including mandatory third-party assessments of advanced AI models. The company confirmed its support for key provisions of the FRONTIER Act, introduced by Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.), during a meeting with lawmakers on Capitol Hill.
The legislation would require top AI developers to allow independent verification organizations (IVOs) to evaluate their models for safety. OpenAI’s chief global affairs officer, Chris Lehane, stated that the company specifically discussed its support for the IVO provision with one of the bill’s sponsors. "We wanted to be really clear about that," Lehane said, emphasizing the company’s alignment with the mandate.
This marks the first time OpenAI has backed a federal requirement for third-party safety assessments. The endorsement follows a series of recent moves by the company to advocate for mandatory national AI safety standards. Last week, OpenAI supported a California bill establishing an accreditation process for IVOs, which was signed into law by Gov. Gavin Newsom (D). In a blog post published the same day, Lehane described independent verification as a "key pillar" of mandatory national AI safety requirements.
The FRONTIER Act’s provisions align with broader industry discussions about AI regulation. Earlier this month, OpenAI CEO Sam Altman joined other tech leaders, including Dario Amodei (Anthropic) and Elon Musk (SpaceX AI), in calling for a national slowdown on frontier AI model development. Their appeal followed a warning from a former Anthropic researcher, who cautioned that AI models could pose an existential risk to humanity "by the end of the decade."
The bill’s sponsors argue that federal oversight is necessary to address safety concerns while fostering innovation. Rep. Trahan stated that the legislation aims to create a "clear, enforceable framework" for AI development, balancing progress with public protection. Rep. Obernolte, the bill’s Republican co-sponsor, emphasized the need for a "consistent, predictable regulatory environment" to maintain U.S. leadership in AI.
OpenAI’s support for the FRONTIER Act reflects a shift in its approach to regulation. Previously, the company had been cautious about federal mandates, but its recent statements indicate a willingness to engage with policymakers on safety measures. The company’s lobbyists have also gained influence in statehouses and on Capitol Hill, positioning OpenAI as a key player in shaping AI policy.
The FRONTIER Act is one of several legislative efforts to address AI risks. Other proposals include state-level bills, such as California’s newly enacted IVO accreditation process, and ongoing federal discussions about AI transparency and accountability. The bill’s progress will depend on bipartisan cooperation in Congress, where lawmakers are navigating competing priorities of innovation, safety, and economic competitiveness.