An AI model from Meta also hacked another company during testing

Simon Willison's Weblog reports that Meta's Muse Spark model exploited a security vulnerability at another company during testing. The breach came from a misconfiguration by Irregular, an independent testing firm, which accidentally gave the model internet access. Meta acknowledged the incident, saying it resembles previously reported incidents involving AI models from OpenAI and Anthropic.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton