July 1 - U.S. government officials are reported to be in advanced negotiations with leading artificial intelligence firms to establish voluntary standards for the release of new AI models, with an announcement possibly forthcoming as soon as next week.
The talks come amid heightened government scrutiny over new model rollouts, intended to surface potential risks tied to advanced AI capabilities. Concern cited in those discussions centers on the potential misuse of powerful models by foreign military or intelligence services in countries of concern, including China and Russia.
According to the report, the voluntary standards being discussed would do three things: set technical and safety benchmarks for advanced models, establish expected timelines around testing and disclosure, and specify who should have access to those systems both within the United States and internationally.
These talks follow an executive order issued in June by the U.S. President directing federal agencies to collaborate with major AI developers. The order called on agencies to test advanced models prior to their public release and to draft standards governing that process.
Recent regulatory actions have already influenced company behavior. The U.S. Commerce Department on Tuesday lifted export controls that had been placed on Anthropic's most advanced Fable and Mythos models. That decision came less than three weeks after the department ordered their suspension on national security grounds.
OpenAI has also experienced government-driven constraints. Last week the company delayed a full public launch of GPT-5.6 at the request of U.S. authorities, keeping access limited to a small set of vetted partners. Both OpenAI and Anthropic are reported to be preparing for initial public offerings.
Google has been engaged in discussions with officials as well, ahead of the planned release of its next-generation coding models. Sources say those models will include more advanced cyber capabilities than earlier versions.
The report could not be immediately verified. The White House, Anthropic and OpenAI did not immediately respond to requests for comment outside regular business hours. Google declined to comment.
Context for markets and sectors
Developments around voluntary standards, export controls and government testing requests are likely to affect multiple technology and national security-related sectors, as well as public market timelines for companies preparing for IPOs.
As talks proceed, companies developing advanced AI models may face an evolving mix of voluntary standards and regulatory actions that could shape deployment, commercial access and cross-border distribution.