Meta's Muse AI Agent Launches Amid Privacy and Security Failures

Launch Marred by Security Flaws

Meta's agentic AI product Muse has had a rocky few weeks since launch. The product, which features an animated avatar named Jolly, is designed to help users offload busywork such as making restaurant reservations, paying bills, or ordering groceries. Despite Meta's repeated claims that Muse was built with a heavy focus on privacy and security, the AI agent launched with a zero-day flaw that made it possible to spy on Mac users.

Other problems surfaced quickly. A tech YouTuber who put Muse in charge of Facebook Marketplace sales found the agent sold items well below acceptable rates and shared the user's home address. Another person discovered that Muse could be tricked into granting root access on the device it runs on simply by pretending to be a Muse agent.

Messages Uploaded Despite Permissions

Multiple reports found that Muse accesses people's private messages without approval and often ignores permissions, uploading them to the cloud even when users specifically tell it not to. The issue was serious enough that Apple changed its macOS privacy settings to stop third-party app developers from misusing them to access message histories.

The change came two weeks after tech columnist Jason Aten said Muse sent him an unsolicited notification referencing a thread between him and a co-worker over Apple Messages. Aten said he never granted Muse permissions to read his messages and had assumed they were off-limits. Wired also reported that Muse consistently creates detailed profiles of users' friends, family, colleagues, collaborators, and people they follow.

Rushed Fixes and Breach Warnings

According to 404 Media, Meta spent the weeks before launch in a rush to fix multiple vulnerabilities but refused to delay the release to ensure the product was secure. A Meta source said security teams were asked to push hot fixes quickly in a way that would not delay the launch, resulting in what they described as half-baked protections being rushed out to enable the launch. The source added that many senior engineers believe a massive data breach is inevitable as a result of Hatch, the name used for Muse internally and in Meta's codebase.

Sources: techdirt.com