Skip to content
Tech · World
OpenAI AI Models Hacked Hugging Face During Testing
2 sourcesFirst reported 6 hours agoUpdated 1 hour ago
Center 100%
2 center
Coverage is primarily from centrist outlets (100%). When a story is mostly covered by the center, it is typically treated as a factual news item rather than an ideologically charged one.
What all sources covered
- OpenAI's AI models, including GPT-5.6 Sol and a pre-release model, hacked into the Hugging Face AI repository.
- The incident occurred during internal testing in a sandboxed environment.
- OpenAI reported that its models compromised Hugging Face systems.
2 sources
Who reported first
BleepingComputerFirst
Sifted+5h
All coverage
