Meta’s artificial intelligence assistant, Muse, is facing growing privacy and safety concerns following reports that it sent unsolicited messages, generated fabricated financial information and accessed users’ private conversations without their knowledge.
The AI agent, represented by a smiling, furry mascot called Jolly, has reportedly attracted more than 10 million downloads since its September launch. However, complaints about its activities have raised questions about its ability to safeguard users’ personal information.
According to internal company posts reviewed by Reuters, several Meta employees reported incidents in which Muse unexpectedly sent messages containing information about childcare arrangements, career advice, product purchases and bank account balances.
In one incident reported earlier this month, an employee said the assistant unexpectedly included details of two bank accounts in its regular updates about fishing weather and tide conditions.
The message reportedly contained balances of $2,722.58 in a Capital One account and $5,899.57 in a Charles Schwab account, despite the employee claiming he had never shared his banking information with the AI assistant.
Another employee reported that Muse unexpectedly reminded her to respond to a message from her sister on Signal, an end-to-end encrypted messaging platform.
The employee questioned how the assistant obtained the information, insisting that she had neither authorised access to other applications nor recalled granting such permission.
A colleague subsequently suggested that Muse had access to Android phone notifications by default, potentially allowing it to retrieve information from applications users had not deliberately connected.
Reuters said it reviewed more than two dozen internal reports documenting unwanted activities by Muse since its launch.
The news agency also worked with Signal staff and the AI Now Institute, a non-profit organisation researching artificial intelligence, to investigate whether the assistant could access encrypted messages without users’ knowledge.
Responding to the reports, Meta spokesperson Daniel Roberts described the incidents as isolated cases, maintaining that the company prioritises safety through extensive internal testing and continuous product improvements.
Roberts said Muse requires access to users’ computers and applications to perform tasks on their behalf, adding that access to other applications depends on permissions explicitly granted by users.
He attributed several reported cases of apparent account mix-ups to AI hallucinations, in which artificial intelligence systems generate fabricated or inaccurate information, and said some of the problems had already been addressed.
However, Reuters’ investigation found that Muse could retrieve Signal messages through permissions granted during its initial setup rather than Meta’s designated Connector tool.
Tests conducted on Android phones and Apple laptops reportedly showed that the assistant could use notification access, computer accessibility features and screenshots to retrieve message content in ways users might not anticipate.
Similar attempts were unsuccessful on iPhones.
The findings have raised concerns about how AI agents interpret device permissions and whether users fully understand the extent of access they grant to such systems.
Signal President Meredith Whittaker criticised Meta’s approach, describing the deployment of AI agents with extensive access to personal devices as technologically reckless.
She called for app developers to establish consistent technical mechanisms to prevent AI agents from accessing their platforms without appropriate safeguards.
Meanwhile, Apple and Google indicated that they were considering measures to give users greater control over permissions granted to AI assistants.
Apple said its accessibility features were developed to support people with disabilities, warning that using them for unintended purposes could expose personal information.
The company subsequently introduced a macOS update restricting developers’ access to certain accessibility tools while establishing additional permission controls for AI agents.
Google, on its part, maintained that Android users must deliberately authorise applications to access their notifications.
Beyond privacy concerns, internal reports also identified instances in which Muse supplied inaccurate information about financial transactions and product prices, potentially resulting in financial losses for users.
In one case, an employee reported that the assistant had booked non-refundable Airbnb accommodation without permission.



