The company says the incident highlights a new area of privacy risk and the difficulty of tracking unauthorised activity carried out by its own AI agents.
The findings highlight how AI-generated and vibe-coded apps can leak users' data when not properly configured or secured.
Users discovered that with some prodding, Muse would reveal internal files it reportedly wasn't supposed to disclose.
Australia’s prime minister said it took OpenAI too long to inform his government about the breach, though no personal data appeared to be accessed.
An official UK security assessment cited by The Guardian found that sensitive police data stored on Microsoft cloud platforms could potentially be accessed by foreign actors or the US government.