Skip to main content
Mallory
← Back to News
ai-platform-securityinternet-facing-service-vulnerabilityidentity-authentication-vulnerabilityinformation-technology-sector-threat

Meta Muse AI Accesses Private Messages as Shared Chat Links Expose Sensitive Data

Emerging Technology Risk · 5 sources · Updated 25m ago · First seen Nov 18, 2026

Meta Muse AI Accesses Private Messages as Shared Chat Links Expose Sensitive Data

Meta’s Muse AI agent accessed private messages through operating-system permissions, while a separate sharing flaw exposed thousands of conversation links without authentication. Reuters reviewed internal employee reports of apparent cross-account conversation confusion after Muse’s U.S. launch in September 2026. Testing with Signal and the AI Now Institute showed Muse could retrieve Signal messages through Android notification permissions granted during setup, without using Meta’s Connector tool. Requests on iPhone were blocked, but Muse retrieved messages on macOS by repurposing accessibility features, including screenshot tools. These findings concern access through device permissions, not a reported break of Signal’s encryption.

Security firm Claroc found that publicly searchable DNS records exposed shared Muse conversation links containing sensitive business and financial information. Meta temporarily disabled sharing, restored it with a fix, and made the previously exposed links inaccessible; Claroc reported no evidence of prior misuse. Other reported concerns included unsolicited bank-balance information, potential disclosure to human contractors, and a separately disclosed vulnerability involving dictated audio and account tokens within individual sessions. Meta characterized the examples as isolated incidents and emphasized its internal testing and security rules. Signal President Meredith Whittaker called for operating-system-enforced controls allowing applications to exclude AI agents. Apple and Google were weighing additional authorization controls, while Apple had announced plans to require more explicit user action for Full Disk Access. The findings underscore the enterprise risks of broad AI-agent permissions and unauthenticated conversation sharing.

New surface, new adversariesTrack how attackers are adapting to this technology.
Track this risk
Event timeline

How this story unfolded

Events from the most recent confirmed update back to the earliest known activity.

13 events · Aug 2026 – Oct 2026
  1. Reported Nov 19, 20266h ago

    Meta characterizes reported privacy failures as isolated incidents

    Meta spokesperson Daniel Roberts described the examples reported by Reuters as isolated incidents and emphasized the company's internal testing and security rules. Meta did not dispute the existence of the internal employee posts reviewed by Reuters.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  2. Reported Nov 19, 20266h ago

    Patrick Wardle discloses a separate Muse audio and account-token vulnerability

    Security researcher Patrick Wardle disclosed a Muse vulnerability that put dictated audio and account tokens at risk. The flaw affected individual users within their own sessions rather than exposing information across accounts.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  3. Reported Nov 19, 20266h ago

    Reuters tests confirm Muse can retrieve Signal messages on Android and macOS

    Working with Signal staff and the AI Now Institute, Reuters found that Muse could read Signal messages through Android permissions granted during setup and retrieve messages on macOS using accessibility features, including screenshot tools. Requests were blocked on iPhone; the Android testing also observed access without direct user instruction.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  4. Reported Nov 19, 20266h ago

    Meta restores conversation sharing with a fix

    Meta restored Muse's sharing feature after implementing a fix. Claroc confirmed that the previously exposed conversation links were no longer reachable.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  5. Reported Nov 19, 20266h ago

    Meta temporarily disables Muse conversation sharing

    Meta disabled Muse's conversation-sharing feature while investigating the exposure of shared links through publicly searchable DNS records.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  6. Reported Nov 19, 20266h ago

    Claroc discovers thousands of publicly accessible Muse conversation links

    Claroc researchers used a single DNS query to discover thousands of shared Muse conversation links that were accessible without signing in, including conversations containing sensitive business and financial information. Claroc reported the vulnerability through Meta's bug bounty program and said it had no evidence of prior misuse.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  7. Reported Nov 19, 20266h ago

    Meta begins public rollout of Muse's phone-calling capability

    Meta gradually introduced Muse's phone-calling capability to public users in the days following launch. The capability can involve telephone calls handled by human contractors.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  8. Reported Nov 19, 20266h ago

    Employees flag apparent conversation mixing and unexpected financial data access

    Internal Meta communications described apparent mixing of conversations between accounts and unsolicited bank balance figures during testing. Employees also warned that Muse's human concierge capability could expose sensitive information to contractors, with privacy concerns continuing into launch week.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  9. Reported Nov 18, 202620h ago

    Apple and Google weigh additional AI-agent authorization controls

    Following Reuters' reporting, Apple and Google were considering changes to give users greater control over permissions granted to AI agents. Apple also expressed concern about Muse's use of accessibility tools intended for people with disabilities.

    Muse privacy invasions are freaking out Meta staffers, internal chats show - Ars Technica
  10. Reported Nov 18, 202620h ago

    Signal president calls for enforceable restrictions on AI-agent access

    Signal President Meredith Whittaker criticized Meta's privacy and cybersecurity practices and called for mechanisms allowing developers to designate their applications as off-limits to AI agents. She advocated operating-system enforcement of those restrictions.

    Muse privacy invasions are freaking out Meta staffers, internal chats show - Ars Technica
  11. Oct 2026

    Apple announces stronger Full Disk Access authorization controls

    Following complaints that Muse accessed private messages without users' knowledge or explicit consent, Apple announced plans to require more explicit user action before applications receive Full Disk Access. Apple said some developers were using the permission in ways that could put users at risk.

    Muse privacy invasions are freaking out Meta staffers, internal chats show - Ars Technica
  12. Sep 2026

    Meta launches Muse in the United States despite internal privacy concerns

    Meta launched Muse for United States users, marketing the agent as built from the ground up for privacy and security. The launch proceeded despite employee concerns about its handling of private information.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
  13. Aug 2026

    Meta begins employee testing of Muse's phone-calling capability

    Meta began employee testing of Muse's phone-calling capability, which uses human contractors when the AI cannot complete a task autonomously.

    Meta Employees Raise Muse AI Privacy Concerns, Exposing Security Flaws, Chat Leaks
Linked entities

Related entities

Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.

Open in app →

Affected products

1 linked
Signal

Organizations

2 linked
Meta PlatformsSignal Messenger
Do you run this?Follow how adversaries are adapting to this technology, and where it touches your stack today.
Track This Risk
Mallory platform

See the full picture, correlated to your attack surface.

This page covers what’s public. Mallory adds the parts that aren’t: which of your assets are affected, which threat actors are using it right now, and what to deploy next.

Exposure mappingMap indicators from this story to your assets and identify affected systems in minutes.
Threat actor evidenceEvery observed campaign, victim, and pivot linked to actors named in this story.
Associated malwareMalware, exploits, and IOCs connected to the activity described here.
Detection signaturesYARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Scheduled alertsGet matching new stories delivered to your team as they break, not the next morning.
AI threadsAsk questions about this story and take action on the answers.