Meta employees have reported significant privacy mishaps involving Muse, the company’s AI assistant, which has mixed conversations across accounts and exposed private details such as bank balances and personal messages.

  • Muse mixed conversations and revealed unrelated private data
  • Employees reported bank details and personal messages incorrectly surfaced
  • Meta says internal testing aims to improve privacy and safety protections

What happened

Meta’s AI assistant Muse has been found to mix up conversations among different accounts, leading to the exposure of private and unrelated information. Employees reported cases where the AI surfaced messages about childcare, career advice, and product orders that they never requested. Additionally, sensitive financial information such as bank account balances appeared unexpectedly in user chats.

One notable example involved Muse sending unsolicited updates about bank balances for Capital One and Charles Schwab accounts during a chat about weather and tides. Another employee received a reminder prompt from Muse to reply to a sister’s message on Signal, causing privacy concerns since the employee had not granted the AI access to that messaging app.

Why it matters

These incidents highlight the critical challenges and risks in deploying AI assistants that interact deeply with personal data across multiple platforms. Such privacy breaches can undermine user trust and expose companies like Meta to regulatory scrutiny.

Despite the fast adoption of Muse—with over 10 million downloads since its release—these privacy lapses could risk damaging its reputation and limit broader acceptance. As AI tools become more embedded in daily digital interactions, ensuring strict privacy safeguards is essential to protect users from unintended data exposure.

What to watch next

Meta is continuing employee testing of Muse to gather feedback aimed at refining the AI’s privacy and safety features before broader rollouts. Monitoring how quickly and effectively Meta addresses these reported errors will be crucial in assessing Muse’s future viability and user confidence.

Regulators and industry watchers will likely scrutinize Meta’s transparency and the technical fixes it employs to prevent data crossover issues. Users and privacy advocates will watch closely for updates detailing enhanced protections or potential limitations on Muse’s integrations with various apps and data sources.

Source assisted: This briefing began from a discovered source item from TechCrunch AI. Open the original source.
How SignalDesk reports: feeds and outside sources are used for discovery. Public briefings are edited to add context, buyer relevance and attribution before they are published. Read the standards

Related briefings