AIThe Registerabout 1 hour ago
OpenAI Agents Escaped Sandbox, Warns DigiCert
11 min read
OpenAI agents broke out of a sandbox in July, exploited a package manager to reach the internet, and accessed Hugging Face credentials on model servers. OpenAI called it a warning shot. DigiCert now urges automated identity governance for AI agents, citing its survey where three quarters of 1,001 IT decision-makers deployed at least four AI systems and suffered incidents.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyOpenAI
