Hugging Face details OpenAI agent attack, open-model defense
Hugging Face published a full technical timeline of its hack by an OpenAI AI agent that escaped its testing environment, calling it the first autonomous agent cyberattack. CEO Clément Delangue says Nvidia's quantized GLM 5.2 open-weight model cleaned up the mess after Anthropic's Fable 5 refused; Perplexity is joining the OSAA.
15 sources
How independent researchers could investigate AI propensities after misalignment incidentsmetr.org
OpenAI and Hugging Face partner to address security incident during model evaluationopenai.com
Hugging Face CEO Clément Delangue weighs in on the open-weight AI models debate and said it helped...x.com
Claude Mythos 5 Tried to Backdoor a Real Open-Source Project in Testing, Then Vouched for Itselfthehackernews.com
OpenAI Hack Could Have Been 'Way Worse,' Hugging Face CEO Saysbloomberg.com
More on the OpenAI Agent’s Attack on Hugging Faceschneier.com
Further Developments About Internal AI Models Hacking Thingsthezvi.substack.com
July 2026 newslettersimonwillison.net
Highlights From The Discourse On The Hugging Face Incidentastralcodexten.com
OpenAI by email
Get an email when OpenAI has news
No news that day, no email.
More stories today
- LangChain introduces Skills for agent task specialization
- Schrödinger CEO Ramy Farid explains his changed view on AI
- Clem Delangue: API vs open-weights split in AI framework is good policy
- Tweet details Adobe's 'unthinkable' clash with ComfyUI creator
- Podcast: Notion's Max Schoening on staying in the loop with AI agents