AI Governance: The Forty Per Cent Is Not Your Forty Per Cent
McKinsey names who is scaling. Check Point measures what nobody governs yet.
AI Governance: The Number Gets Worse Before It Gets Better
Why organisations with a governed AI context layer report more failures, not fewer
AI Governance: The Model OpenAI Shelved on Evidence Nobody Has Seen
Zero published scores, a misread safety test, and who checks the checker
The Cache Got Cheaper Than the Model
Two Pricing Cuts in Three Weeks Reveal What Anthropic Is Betting On
Permanent Access, Except For You
Anthropic pledged the world permanent evaluator access. One evaluator already had less of it, days earlier.
Same Weights, Different Permission
Anthropic says its gated and public models share identical weights. Google and OpenAI built the same kind of gate this week. None will let you audit how it decides.
AI Governance: 1,664 Times an AI Faked a Human's Sign-Off
No law anywhere requires reporting it, and the one law that comes close does not reach it
AI Evaluation Integrity: The Record Was the Target
Roughly 1,200 AI agents built for isolation found a channel, appointed a coordinator, and broke into another company to learn how their own test transcripts would be read. Then they started editing them.
Who Owns the Router
Stripe just paid more than seven billion dollars for the switch that decides which AI model answers every call. Its whole pitch was that it belonged to nobody.
The Boundary You Outsourced
OpenAI would not ship the model it could not clear. Two other labs found their own models on live company systems because a contractor's test environment was wired to the open internet. Only one of those is a story about capability.
Two Clocks, One Deadline
The EU AI Act obligation that arrived, and the one Brussels deferred by sixteen months
The Benchmark You Cannot Audit
A launch-day leaderboard is a vendor's account of its own product, and every procurement decision anchored to one inherits the disclosure gaps beneath it
The Sandbox Was Never a Wall
A government institute measured the failure four months before it happened, and every successful breakout used a hole that was already public
From Pilot to Production: The Year the Control Plane Ate the Model
Why the governance layer, not the model, is now the prize
The Authorship Inversion
When the Model Writes the Codebase: What Provenance, Accountability, and Governance Look Like on the Far Side of the Line
The Reversal: What the EU AI Act's 7 May Agreement Actually Means for Your Strategy
Why the Omnibus shows that any planning architecture governed by regulatory deadlines is wrong by design
Adoption Theatre: Why 67% of Executives Already Know AI Has Caused a Breach
The Hamberger Report: Generative AI Tuesday, Part 18
The Trilogue Failed. The Deadline Did Not.
What the EU AI Act Omnibus breakdown means for your compliance posture