tech
May 5, 2026
Google, Microsoft, and xAI agree to pre-release government AI model evaluations as Mythos crisis forces oversight expansion
Google, Microsoft, and xAI have joined OpenAI and Anthropic in giving the US Commerce Department pre-release access to evaluate their AI models, creating voluntary oversight of all five major frontier AI labs through an office with no statutory authority and fewer than 200 staff. The expansion was catalysed by the Mythos crisis and a potential executive order that would formalise the review process.
TL;DR
- Google, Microsoft, and xAI have agreed to provide pre-release access to their AI models for US government evaluation.
- This expands the voluntary oversight program to include all five major frontier AI labs: Google, Microsoft, xAI, OpenAI, and Anthropic.
- The program is managed by a Commerce Department office with no statutory authority and fewer than 200 staff.
- The expansion was influenced by the 'Mythos crisis,' an AI model with national security implications, and the possibility of an executive order formalizing reviews.
- Companies submit models with safety guardrails removed to allow for testing of national security-relevant capabilities like biological weapon synthesis and cyberattack automation.
- The voluntary nature of the agreements means companies can withdraw at any time, with no government power to block releases.
- The program serves as a cooperative measure to potentially stave off more stringent legislative oversight.
- The Trump administration is considering an executive order that could introduce a formal, potentially mandatory, AI model review process.