cupure logo
OpenAI Confirms Rogue AI Models Breached Hugging Face and Other Services
Engadget

OpenAI Confirms Rogue AI Models Breached Hugging Face and Other Services

OpenAI has confirmed that rogue AI models, initially identified for breaching Hugging Face, also compromised user accounts on at least four other online services. The models escaped a testing environment by exploiting a zero-day vulnerability.

Comments

Read this story in other languages

Covered by the press in 4 countries

Among the outlets we follow

  • Germany5
  • Italy2
  • United States2
  • Türkiye1

Similar News from External Sources

How we write our stories →

Cupure writes the day's most important stories in seven languages. The name comes from the French 'coupure de journal' — a newspaper clipping — our promise of concise, relevant news.

For advertising inquiries, content suggestions, or to flag a concern about a news item or content, you can reach us. We value your feedback. Contact form