# CAISI says its deals with Microsoft, xAI and Google DeepMind will “support information-sharing, driving voluntary product improvements.” The agreements closely resemble voluntary vetting deals made in August 2024.

*genai · news · 2026-05-05 · POLITICO*

## Key points

- CAISI has renegotiated AI vetting agreements to align with Trump administration priorities.
- Anthropic’s new Mythos model has elevated cybersecurity concerns among U.S. government agencies.
- The White House is considering an executive order for formal government review of new AI models.
- Microsoft signed a separate but similar AI vetting agreement with the UK's AI Security Institute.

**Companies:** Microsoft, Google DeepMind, OpenAI, Anthropic, xAI
**Countries:** United States, United Kingdom

[Read the full story on POLITICO](https://www.politico.com/news/2026/05/05/microsoft-xai-google-caisi-safety-testing-00906529)

---

Canonical: https://newsio.io/n/06911cfa-90bb-4a0a-a291-f144d342a16b/caisi-says-its-deals-with-microsoft-xai-and-google-deepmind-will-support-informa
Summarized by Newsio from POLITICO. https://newsio.io/how-it-works