The Trump administration is considering a possible executive order to introduce pre-release oversight of advanced artificial intelligence models, according to multiple reports citing the New York Times and confirming details from other outlets. The proposal would create a working group made up of government officials and technology executives to examine review procedures for new AI models before they are publicly released. Bloomberg, Forbes, CNA, The Hindu, Engadget, and CNBC all report that the White House is exploring structured oversight rather than a purely hands-off approach.
The discussions are described as a reversal from earlier policy, when the administration broadly supported rapid development and reduced safety-focused review steps associated with the previous Biden-era approach. Reporting also says the working group would consider various options, including models of oversight used elsewhere, such as the UK’s approach that involves multiple government bodies assessing safety standards.
Some reporting indicates the administration wants to understand potential risks and capabilities, including possible cyber-related concerns, and to determine whether the government should receive early access to frontier models without necessarily delaying or blocking public release. The meetings reportedly include communications with major AI companies such as Anthropic, Google, and OpenAI.
