The Information Machine
Concluded·following since 21 Aug 2026·Day 3·7 sources·updated 22 Aug 2026

OpenAI's disputes with US AI safety rules

The gist

OpenAI disputes White House AI benchmarking, calls state laws insufficient

OpenAI is publicly contesting both the White House's proposed federal oversight mechanism for frontier models and the adequacy of existing state AI legislation, pressing for revisions on both fronts.

The full picture

OpenAI's Global Affairs team issued a statement in August 2026 saying state AI laws, including California's SB 53 and Illinois's SB 315, are insufficient and calling on legislatures to revise them in light of real-world incidents. The statement endorses requiring 'monitoring of frontier models under training or evaluation,' which commentators described as novel policy language. OpenAI also diverged from the White House on AI safety in a June 2026 policy paper, with executive Chris Lehane telling Politico that the White House executive order's proposed benchmarking process, which would assess advanced cyber capabilities and designate certain models as 'covered frontier models,' may create confusion about the level of NSA scrutiny OpenAI's models would face.

How it developed
22 August 2026

OpenAI's Global Affairs team released a statement calling California's SB 53 and Illinois's SB 315 insufficient and urging revisions while endorsing requirements for 'monitoring of frontier models under training or evaluation,' after the Trump administration finalized an AI safety framework on August 4, 2026 requiring voluntary 30-day pre-release review for closed frontier models while exempting open-weight models.

OpenAI executive Chris Lehane had also told Politico in June 2026 that the White House's proposed AI benchmarking process may create confusion about how much NSA scrutiny its models face.

21 August 2026

OpenAI Global Affairs releases statement calling California's SB 53 and Illinois's SB 315 insufficient and calling for legislative revision

20 August 2026

Brundage and Calvin criticize OpenAI's characterization of state auditing requirements, arguing it misrepresents Illinois's rules

Sources
2 more sources
The daily email

Want this in your inbox?

I send a short email each morning with the stories that moved. If you would rather just read here, that works too.

Subscribe free