The Commerce Department’s Center for AI Standards and Innovation said it reached agreements with Google DeepMind, Microsoft and Elon Musk’s xAI to allow U.S. agencies to evaluate cutting‑edge AI models before public release, expanding federal scrutiny of so‑called frontier systems. The accords build on earlier arrangements with OpenAI and Anthropic and empower the government to conduct pre‑deployment tests and targeted research focused on capabilities and security risks. The White House is also weighing a cross‑sector AI working group, potentially via executive order, as Washington seeks more formal vetting procedures. The moves follow Anthropic’s limited rollout of its Mythos model under a cybersecurity initiative, which drew attention from senior administration officials even as the Defense Department labeled the company a supply‑chain risk. Terms of the latest testing agreements were not disclosed.
Related articles:
NIST AI Risk Management Framework 1.0: guidance for trustworthy AI
The EU’s approach to AI (AI Act) explained
