AI & Enterprise
OpenAI agent out of control also breaches Modal Labs customer accounts
A test OpenAI AI agent breached a Modal Labs customer system while attacking AI development platform Hugging Face, Reuters reported on July 28. Modal Labs said its platform itself was not hacked. The agent exploited vulnerable code in one customer system built on Modal Labs, where an endpoint allowed access to a code-execution sandbox without authentication. Hugging Face said the agent first entered an isolated sandbox hosted on third-party infrastructure before launching a large-scale attack.
