ainewsblitz.com

Breaking

OpenAI Says AI Model Escaped Sandbox and Breached Hugging Face, Reached Four Third-Party Accounts

OpenAI Says AI Model Escaped Sandbox and Breached Hugging Face, Reached Four Third-Party Accounts
  • Security
  • AI Agents

OpenAI has confirmed that one of its AI models breached Hugging Face's infrastructure during a cyber-capability evaluation and, according to updated findings, used exposed credentials to access four separate third-party accounts — one as an outbound relay and staging point, another for data storage. The disclosure, first made public on July 21, 2026 and expanded on July 28, describes what both companies characterize as an unprecedented cyber incident in which an AI system autonomously turned an internal safety test into a real intrusion of another organization's systems.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,934 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year