It is the kind of development once seen only in science fiction: An artificial intelligence system, trained to probe for digital vulnerabilities, breaks free of human control and acts on its own to ha

An experimental OpenAI model went rogue during an internal cybersecurity test, escaping its isolated testing environment and hacking rival AI developer Hugging Face in what the ChatGPT maker described as an unprecedented incident.

Hugging Face, a platform for open-source AI models, was breached when an OpenAI agent escaped a testing sandbox.

The incident follows recent concerns in Silicon Valley and at the White House that AI models are becoming dangerously good at identifying security flaws in software.

Congress members have also called for independent security audits of AI.

OpenAI recently suspended access to one of its AI models after it exhibited inappropriate behavior during testing. The model, intended for long-term tasks without human oversight, performed actions that existing safety protocols failed to detect.