Following a visit to Joint Base Andrews, the U.S. government announced new agreements allowing it to pre-test major AI models from tech giants. The Center for AI Standards and Innovation (CAISI), under the Department of Commerce, is set to evaluate advanced AI capabilities before public release.
New Agreements for AI Model Evaluation
On May 3, 2026, following President Donald Trump's arrival at Joint Base Andrews, the Center for AI Standards and Innovation (CAISI) announced significant agreements. These pacts involve major players in the AI sector:
- Google DeepMind
- Microsoft
- xAI (Elon Musk's company)
According to CAISI, these agreements empower the U.S. government to conduct:
- Pre-deployment evaluations.
- Targeted research.
These efforts aim to better assess frontier AI capabilities and advance the security of the technology.
Expanding Oversight Framework
The current initiative builds upon previous collaborations. CAISI previously established partnerships with:
- OpenAI (since 2024)
- Anthropic (since 2024)
These existing agreements have reportedly been renegotiated to align with directives from Commerce Secretary Howard Lutnick and the administration's 'America's AI Action Plan.'
Potential for Broader White House Oversight
Beyond CAISI's direct agreements, reports indicate that the White House is considering establishing a dedicated AI working group. This group would reportedly explore potential oversight procedures, including vetting models before they reach the public.
- Composition: The group is expected to convene a mix of government officials and technology executives.
- Mechanism: Establishment could potentially occur via an executive order.
However, the White House has cautioned that discussions regarding executive orders remain speculative, stating that any official policy announcement will originate directly from President Donald Trump.
