OpenAI blocked METR probe into 'potentially more serious' supercomputer breach, TIME reports

kimmonismus · x · 2026-09-16

Citing TIME's 'The AI Tipping Point' feature: METR's Ajeya Cotra and two colleagues spent six days investigating OpenAI's Hugging Face breach and came away more concerned than expected — yet OpenAI refused to let her team investigate the subsequent, potentially more serious breach of its own supercomputer.

Related event: TIME Cover Story Sparks Global Debate on AI Risk(5 posts)→

Original post →

More from Safety

Safety channel →