Superintelligence: Paths, Dangers, Strategies · chapter 14 · id superintelligence-c14-04-unilateral-action-by-a-single-

“Unilateral action by a single AI developer racing ahead without safety precautions poses catastrophic risk.”

needs contextconfidence: high⚠ extracted by pipeline, re-audit pending

Receipts

CNBC: OpenAI Safety Team Dissolutionsource alive

OpenAI dissolved its safety-focused Superalignment team and reallocated safety compute to product development.

CNBC: OpenAI AGI Readiness Head Resigns (October 2024)source alive

Miles Brundage, OpenAI's head advisor for AGI Readiness, resigned in October 2024 -- a further erosion of the safety leadership.

Bletchley Declaration (28 nations recognize the risk)source alive

28 nations signed the declaration explicitly recognizing frontier AI risks and the need for coordinated governance -- an implicit acknowledgment of unilateral racing risks.

This claim is a stable, citable object. If you can falsify a verdict, tell us — corrections are loud here.