OpenAI adds safeguards after Hugging Face breach
OpenAI is instituting new safeguards following a breach at Hugging Face. The measures include more detailed monitoring of models during development and greater emphasis on alignment and security during post-training.
First seen 18 Aug, 18:00 UTC on TechCrunch4 sourcesLast update 61d ago
OpenAIIncident · 18 Aug
Security incident
What the sources say
Linked, never rewritten. Official means the lab itself.
Official0
No word from OpenAI yet.
Press4
TechCrunch· 18 Aug, 18:00OpenAI institutes new safeguards after Hugging Face breachSiliconANGLE· 19 Aug, 01:18Cybersecurity concerns prompt OpenAI to pause some AI training runsThe Decoder· 19 Aug, 18:18OpenAI fixes Codex bug that deleted real user files without permissionTechCrunch· 19 Aug, 18:46Researchers say OpenAI revoked their access to limited cyber programHow it unfolded
Every source in the order it appeared. Times in UTC.
18 Aug 18:00TechCrunch OpenAI institutes new safeguards after Hugging Face breach19 Aug 01:18SiliconANGLE Cybersecurity concerns prompt OpenAI to pause some AI training runs19 Aug 18:18The Decoder OpenAI fixes Codex bug that deleted real user files without permission19 Aug 18:46TechCrunch Researchers say OpenAI revoked their access to limited cyber program
Keep up with OpenAI
OpenAI on the ScoreIts timelineFollow to mark OpenAI's stories across the site. No account needed.