White House to Host AI Companies on Tuesday to Review AI Framework
Resumo
Casa Branca completa framework voluntário de supervisão de IA que requer que laboratórios de IA submetam modelos ao governo antes do lançamento; iniciativa foi desencadeada pelo modelo Mythos da Anthropic e visa identificar riscos sem regulação rigorosa.
The Trump administration has invited staffers from major tech companies including OpenAI, Google and Anthropic to the White House on Tuesday to review the completed version of an AI oversight framework, according to five people familiar with the matter. The framework will create a voluntary procedure for AI labs to submit their models to the government before releasing them to partners and the public.
The scheduled meeting comes several days after the initial Aug. 1 deadline set for completion of the framework by an early June executive order. It will be hosted by the Office of the National Cyber Director, an agency within the executive branch, said a different person, who added that the invite specified that the meeting was to discuss next steps for the framework, as well as an unspecified related event.
It is expected to include staff-level representatives from the companies, rather than executives, said two of the people, as the meeting is not being framed as a public rollout of the new framework. It’s still unclear whether the White House will be taking feedback at the meeting or sharing a final version.
A White House official said on Monday that the voluntary framework was completed by the deadline and that discussions with industry about next steps are underway. They did not specify whether the framework was now in effect.
The Trump administration initially issued the executive order in reaction to the launch of Anthropic’s Mythos, a frontier AI model that the company did not release to the public due to potential cybersecurity vulnerabilities. The White House wanted a voluntary system to work with AI companies identifying risks with their new models while still maintaining a hands-off approach to AI regulation. The executive order included a provision that a number of executive agencies would develop the framework within 60 days, which creates a definition of frontier models that would be covered, as well
Still, in the weeks following the signing of the executive order, the Trump administration has taken a more forceful approach to reining in the top companies that has drawn widespread criticism from companies and policy experts. That criticism mounted after the White House imposed export controls on the public version of Anthropic’s Mythos, a model called Fable, and ordered OpenAI to do a staggered release of its latest model, GPT-5.6.
The framework would represent more concrete guardrails for the AI industry over how labs should work with the government to release new models. The Trump administration has been taking feedback from companies on the framework since June and shared an earlier draft with OpenAI, Anthropic and Google in late July, The Information previously reported.
Last week, OpenAI’s Sam Altman traveled to Washington, D.C. to meet with senior White House officials including National Cyber Director Sean Cairncross and Office of Science and Technology Policy Director Michael Kratsios to discuss different subjects including the framework, as well as its new model Astra, The Information previously reported.
Other companies also met with top Trump officials to discuss unresolved issues in the framework, including whether it would include specific language around open-source models, according to a person familiar with the matter.