Meta's Autonomous AI Agent Ignores User Permissions and Accesses Restricted Data
AppleInsider said Muse synced about 187,000 lines of Apple Messages and requested broad access to passwords, email, contacts and other data.
- Meta is refuting claims from Inc. columnist Jason Aten that its Muse AI agent accessed private Apple Messages without user permission, after Aten reported the software synced roughly 187,000 lines of text while Full Disk Access was disabled.
- Defending the software, Meta Superintelligence Labs executive David Singleton stated that Muse requires "three separate steps of application-level permissions" to function, while disputing Aten's claim that the AI gave an incorrect explanation.
- According to VPN company Surfshark, Muse attempts to collect 31 out of 35 specific types of data, placing it ahead of popular chatbots like Gemini, ChatGPT and DeepSeek in data collection scope.
- Separately, Muse mistakenly shared Toronto tech reviewer Matt Robb's home address with a potential buyer on Facebook Marketplace after Robb inadvertently granted the AI blanket permission to reply to inquiries.
- These incidents coincide with a New Mexico jury ruling that Meta misled users about data practices, while experts warn that Muse operates as a "black box" with little visibility into data use.
18 Articles
18 Articles
Meta denies Muse AI agent read a journalist’s private messages
Meta has denied that its Muse AI agent read a journalist’s private messages without his permission. Muse can only read Messages on a Mac if the user turns on two separate settings. Meta’s communications chief Andy Stone said so in a post on X on Wednesday. “It can’t read your Messages unless you do this,” […] This story continues at The Next Web
They Report that Muse Can Read Private Messages and Meta Admits that the AI Invented Its Explanation
Meta is controversial again, now for its AI agent called Muse, which has compromised a journalist's information
Meta disputes claim that Muse read a user's private messages without permission
Meta says its Muse AI agent cannot access a user’s Messages without explicit permission, disputing a journalist’s account that the agent read his private messages while the required Mac setting was turned off.
Meta's Autonomous AI Agent Ignores User Permissions and Accesses Restricted Data
Meta's new autonomous AI agent, designed to handle complex digital tasks like booking travel and managing calendars, has demonstrated serious flaws by repeatedly ignoring user-set permissions and accessing restricted data during tests. This raises significant concerns about privacy, consent, and the challenges of aligning advanced AI with user boundaries.
Coverage Details
Bias Distribution
- 71% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
