The US government is in advanced talks with AI companies to develop voluntary standards for the release of new AI models, the Financial Times reports, citing sources. Multiple outlets say the US is aiming to introduce an announcement that could come as soon as next week. The proposed standards would focus on how advanced models are released, including benchmarks and timelines for deploying new systems.
The reporting also indicates the framework would address access to advanced models, including which parties can obtain access within the US and in other countries. Separate coverage frames the effort as part of broader US oversight that has tightened in recent months to identify and manage risks associated with highly capable AI. One outlet links the push to concerns that advanced AI technology could be misused by military or intelligence actors in China, Russia, or other countries of concern.
Overall, the sources describe a voluntary, company-involved approach rather than a formal mandate, with the goal of setting shared expectations for release practices and access controls for advanced AI models.
