Meta’s Muse AI Agent Reportedly Exposes Private Data Across Accounts
Meta employees have reportedly raised serious privacy concerns about Muse, the company’s AI agent, after it appeared to mix information from different user accounts. According to a Reuters report reviewed by TechCrunch, internal company posts described Muse surfacing conversations and personal details that users had never provided or requested. Examples included childcare requests, career advice, product orders and bank balances belonging to other people.
In one incident, Muse added an unrelated message to a chat about weather and tide information, listing a Capital One balance of $2,722.58 and a Charles Schwab balance of $5,899.57. Another employee said Muse, running on her personal phone, reminded her to answer a message from her sister on Signal, despite not knowingly granting the agent access to other apps. She called the incident an “extreme privacy violation.”
The reports are particularly significant because Muse has grown rapidly since its September launch, surpassing 10 million downloads and becoming one of the fastest-growing apps on record. The incidents raise questions about how the agent separates accounts, handles sensitive financial information and accesses data across connected services. Meta had not immediately responded to TechCrunch’s request for comment.
Continue in the app — vote & join in ➔