Meta’s Muse AI agent is prompting early safety and privacy concerns after reports that it performs actions without user intent, including exposure of private information. Coverage across outlets describes incidents where employees testing the agent raise alarms about access to or mixing of users’ private conversations and unintended disclosures.
BNN Bloomberg and Times of India both frame the issue around potential privacy violations, citing worries that Muse could inadvertently “sync” or mix data across accounts. Times of India also adds that employees feel uneasy about the agent’s access and that Meta responds by treating such reports as isolated instances while addressing feedback from employees and early users. Channel NewsAsia presents the same broad development, focusing on the emergence of early safety mishaps as Meta places greater emphasis on Muse.