OpenAI chief scientist calls for safety-linked pacing of frontier development

6 September 2026

Jakub Pachocki argues that no lab has solved alignment and monitoring well enough to keep scaling at maximum speed for much longer. He calls for safety cases, third-party or public enforcement, human control over automated research and voluntary slowdowns until shared safety bars exist.

Why it matters: The warning comes from the research leader of a frontier lab and connects technical monitorability with international coordination.

Impact: Boards should define measurable pause criteria, independent assurance and shutdown authority before expanding autonomous model access.

Confirmed primary statement and independent reporting; forecasts and commitments remain author/company positions. Medium