White House briefed Anthropic, Google, and OpenAI on plans for a government AI review process

The buried angle here is sequencing: the White House chose to brief the three dominant frontier labs before any public announcement, which means these companies had advance notice to shape the policy's contours, not just react to them. That's a very different power dynamic than a surprise executive order.

This story lands directly on top of a cluster of coverage from early May. The GPT-5.5 and Mythos cyber parity piece from The Decoder (story 7) is the clearest antecedent: it documented that frontier-grade offensive capabilities were already entering mainstream deployment while Mythos remained restricted, which is precisely the kind of asymmetric risk profile that would prompt a government review mandate. Meanwhile, the Pentagon exclusion of Anthropic (The Verge, May 1) showed that Anthropic's safety posture was already costing it contracts. A pre-release review process could paradoxically rehabilitate Anthropic's government standing by forcing competitors onto a slower, more scrutinized release track, narrowing the gap that the Pentagon deals had opened.

Watch whether the draft executive order includes a defined review timeline with hard deadlines. If the process is open-ended rather than time-boxed, it functions less as a safety gate and more as a discretionary delay mechanism, which would disproportionately disadvantage labs without existing government relationships.