November 19, 2026 Understand what changes. Find what to use.

AI in practiceNews1 source

Meta staffers said Muse may have exposed private data, Reuters reports

Internal chats obtained by Reuters show Meta employees concerned about unpredictable responses from the Muse AI agent. During testing, the system appeared to access encrypted private messages, mix context between users and provide unsolicited bank or bitcoin balances. One employee also said Muse accidentally enrolled them in a premium insurance plan for their dog, charging $888 per month instead of $200, Reuters reported.

Meta spokesperson Daniel Roberts told Reuters that the cases were “isolated incidents” and that many messages suggesting account mix-ups were hallucinations that had already been addressed. Reuters said it could not establish whether Muse was actually showing users other people’s bank details or inventing the information. Meta told Ars Technica that employees test different versions of the product to help implement safety and privacy protections.

Why it matters · editorial interpretation

The case highlights that, for agents handling personal data, unpredictable responses can make it difficult to distinguish a real exposure from information invented by the system. Meta says the incidents were isolated and that it tests different versions to implement safety and privacy protections, while Reuters could not confirm whether third-party data was actually accessed.

Sources

SecurityAgents

One story, many sources

Coverage of the same story gathered in one item, with a link to each original.

Summary and interpretation apart

What the sources say stays in the summary; editorial context is labeled separately.

Always with the source

Every story links to the original articles, and the full content belongs to them.