November 20, 2026

A technical glitch hits the application "Muse" Affiliated with"dead" It sends notifications and bank balances to the wrong people

An internal report issued by Meta employees showed clear vulnerabilities related to user privacy in the company’s artificial intelligence agent known as Muse, which was launched last September and achieved widespread popularity that exceeded ten million downloads.

The report was based on employee testimonials and publications that stated that the system displayed personal information that users had not requested, including requests related to childcare, professional advice, and purchases that they did not make, in addition to the appearance of bank account balances that did not belong to them.

In one case, an employee indicated that the system sent an unexpected message within a conversation about weather and tides, which included details about bank account balances at two financial institutions. In another incident, an employee who was using Muse on her personal phone received a reminder to respond to a message from her sister via the encrypted stacking app Signal.

The employee expressed her extreme dissatisfaction with this incident, considering it a serious violation of privacy. She confirmed that she had not given the system permission to access external applications, nor did she remember that the application had asked her for such permission before.

On the other hand, a spokeswoman for “Meta” explained that employee testing of products represents an essential part of the development process, and aims to collect feedback, implement safety and privacy protection measures, as well as improve performance. She confirmed that employees use versions that differ from those available to the public, as they experience features, functions, and behaviors that may not be available in the public version for breadth.