The Financial Times reports the White House is negotiating with AI firms to
create voluntary standards for new model releases, after U.S. interventions in
deployments by Anthropic and OpenAI. People familiar with the talks say the
standards could be published as early as next week, setting baselines for models
with advanced internet capabilities and establishing release timelines to
streamline future rollouts. The initiative implements President Trump’s AI
executive order from last month; stakeholders are seeking clarity on how the
government will regulate the technology. Officials expect the U.S. AI Standards
and Innovation Center and the NSA to play key roles in drafting and overseeing
the standards. Sources add U.S. officials will specify which entities may access
advanced models domestically and abroad, a step that could form the basis for a
global framework including U.S. allies.
