Working groups from Anthropic, OpenAI and Google have met regularly since July to discuss creation of a voluntary industry AI standards body, according to a September 13 report from The Information. The teams, composed of executives below chief-executive level, have explored safety benchmarks and related rules that the companies could adopt without waiting for a completed government framework.
The talks sit alongside the broader public conversation that followed Dario Amodei’s call to pace frontier development. Sam Altman told an all-hands meeting last week that major AI companies may need to build such a body themselves if government support remains limited.
Key takeaways
- Anthropic, OpenAI and Google working groups have met since July on a standards body.
- The Information reported the discussions on September 13, citing people familiar with the matter.
- The concept centers on voluntary safety benchmarks rather than a new statutory regulator.
What an industry standards body would actually do
According to the report, the discussions focus on safety benchmarks and related operational rules that participants could implement on their own schedules. The structure is framed as private-sector coordination while U.S. government efforts remain incomplete. OpenAI’s Altman has separately said companies should not wait for legislation to begin adopting independent-evaluator access of the kind Amodei proposed.
For the longer content-stack implications of agent and safety infrastructure, see How AI is Reshaping the Content Creator Stack in 2026.
What to watch next
Watch whether any of the three companies publishes a joint or individual statement formalizing the working-group process. Watch whether independent evaluators receive the permanent, employee-level access Amodei described as Anthropic’s immediate next step.