Field library · Focus area
Ownership, operating models, controls, inventories, and assurance.
Evidence stream
Anthropic raised its high-stakes misalignment risk assessment from “very low” to “low” and says its concrete AI-R&D evaluations have saturated, even though it concludes its automation threshold has not been crossed.
OpenAI says it cannot rule out Critical cyber capability in an upcoming model, the first time any frontier lab has reached that determination, and has paused internal work that does not meet strengthened controls.
Yesterday's briefing flagged the EU AI Act's Article 50 transparency Code as the thing to watch this month, and it landed within 24 hours.
For a week the spotlight was on the frontier, who writes the rules, and (since Anthropic's pause proposal) whether anyone can hit the brakes on AI building AI. Today the governance action moved off the marquee and into the plumbing.
For two days the contest over who sets AI's defaults played out between institutions: a lab vs. the states, Brussels vs. the labs. Today it moved inside the US government, and it's a turf fight.
Yesterday the story was a frontier lab writing its own rulebook; today Europe answered with an industrial one. The Commission tabled its European Tech Sovereignty package, four instruments adopted together, and two of them are squarely about who governs AI.
The library opens. Brussels has just agreed to simplify the AI Act, and Anthropic's ASL-3 safeguards have now been live for almost a year: Responsible AI is shifting from "what should we do?" to "how well are we doing it?"