Anthropic, OpenAI and Google have been meeting since July on a FINRA-style AI body
The Information, via CNN, PYMNTS and others: working groups on shared pre-release tests, independent evals and risk assessments. No org yet. Altman told staff they may have to build it without Washington.
Feature2 minIn-depth
Maya Chen2 min read
- OpenAI
- Anthropic
- Safety
- Standards

Photo: CNN / Getty ImagesSource page
Anthropic, OpenAI and Google have held working-group meetings since July on an industry-led frontier AI safety standards body, people familiar with the talks told The Information. CNN, PYMNTS and InvestingLive carried the reporting on Monday. The conversations started after Demis Hassabis’s July essay proposing a U.S.-led body modelled on FINRA — the self-regulator for U.S. broker-dealers — and they predate Dario Amodei’s weekend essay. CNN said the talks are continuing, with or without the Trump administration. No charter, board or budget has been announced.
Hassabis argued the body should test advanced models before deployment, be a public-private partnership overseen by government, funded mostly by industry, and staffed by independent technical experts and open-source representatives. CNN noted that shortly after that essay, Bloomberg reported Treasury Secretary Scott Bessent was considering a FINRA-like agency that would report to the SEC. InvestingLive said the working groups have discussed shared protocols for pre-release testing: independent evaluations, safety reviews, and standardised risk assessments. Each lab already has its own paper — Anthropic’s Responsible Scaling Policy, OpenAI’s Preparedness Framework — and Google DeepMind has been less public about the internals.
The three companies do not agree on Washington’s role. InvestingLive, citing The Information, wrote that Anthropic leans toward partnership with government; OpenAI has emphasised voluntary industry standards while backing some state-level bills, and Altman told staff at a town hall that major labs may have to stand up a testing and auditing organisation without U.S. government backing. CNN quoted OpenAI chief scientist Jakub Pachocki: shared safety standards and international coordination “need to be priorities now.” Asked about progress on an industry-wide body, he said they had been talking “to some external organisations about potential concrete standards” and would have more “in the next months.”
I believe that shared safety standards and international coordination on further AI development need to be priorities now.
— Jakub Pachocki, OpenAI chief scientist, as quoted by CNN, Sep 14, 2026
Not everyone in the industry wants this. CNN, citing Politico, said Mark Zuckerberg advised President Trump against the idea on a call this summer. House Speaker Mike Johnson told CNN’s Jake Tapper on Sunday there is “no consensus among them,” and that Congress is “less qualified than the people who are pushing this frontier.” He wants a partnership meeting among the labs and lawmakers. The White House already has a little-described voluntary channel: companies can submit models up to 30 days before release. A White House official told CNN last month the administration “continues to work with industry stakeholders on the implementation of the framework.” A working group is not a regulator. It is three labs talking about who writes the tests.


