The UK AI Security Institute tested Mythos and acted on its findings while restricting access to eight European cyber agencies, illustrating how frontier AI is reshaping intelligence-sharing relationships among allies. Meanwhile, xAI announced a series of Grok releases—Grok 4.4 at 1 trillion parameters for early May, Grok 4.5 at 1.5 trillion for late May, and Grok 5 positioned as AGI. OpenAI saw executive departures including Bill Peebles, Kevin Weil, and Srinivas Narayanan. The White House directed the War Secretary to release UAP files, and Rep. Ogles cited ultra-classified UAP evidence in public remarks. The full scope of how these developments interconnect remains unclear.
Attorneys should monitor how frontier AI deployment is outpacing formal risk governance. The pattern of continued government reliance on models flagged internally as risky, combined with fragmented international access and executive departures at leading labs, signals that institutional momentum around AI development may be overriding traditional security protocols. Watch for regulatory responses, supply chain restrictions, and whether classified technology disclosures accelerate as AI capabilities advance.