Meta employees have expressed concerns that Muse, the company’s AI agent, is mixing up conversations between different accounts and leaking users’ private data, according to a new report from Reuters that looked at internal posts.
In some of these instances, employees said Muse surfaced messages about child care requests they never made, career advice they never asked for, orders for products they never ordered, and bank account balances that were not theirs.
One employee told a co-worker that Muse randomly sent additional messages in chat about weather and tidal information, saying, “By the way, I have two bank updates. Your balance at Capital One is $2,722.58, and your balance at Charles Schwab is $5,899.57.”
In another example, an employee using an AI agent on his personal phone received a reminder from Muse to respond to a message from his sister on Signal, a popular end-to-end encrypted messaging app.
“This seems like an extreme invasion of privacy. I have never given anyone access to other apps and I don’t remember being asked about this,” the employee said in the post, according to the report.
Released in September, Muse is one of the fastest growing apps on record, with over 10 million downloads.
Meta did not immediately respond to TechCrunch’s request for comment.
