OpenAI backs federal AI safety checks by independent assessors

OpenAI is backing a key part of the bipartisan FRONTIER Act, a House proposal that would require leading AI labs to let independent verification organizations assess whether advanced models are being developed safely.

Politico reported that OpenAI global affairs chief Chris Lehane told reporters the company can support the independent-verification provision after discussing it with one of the bill’s sponsors. The measure from Reps. Jay Obernolte and Lori Trahan is framed as a first federal AI safety framework for the most capable systems, not a broad rulebook for every AI startup.

The move follows OpenAI’s own policy post calling for mandatory, capability-based national AI safety requirements. In that post, the company says independent assessment, cybersecurity protections, incident reporting and shared measures for tracking AI-driven research acceleration should become part of a federal framework.

This matters because frontier AI safety has largely depended on voluntary lab commitments and state-level experiments. OpenAI’s support does not make the bill law, and the details still matter, but it gives Congress a clearer signal that at least one major lab is prepared to accept outside safety checks as mandatory infrastructure rather than a purely voluntary norm.