Case meta-ai-discover-feedModel & data leakJun 12, 2025United States

Meta AI app's Discover feed broadcasts users' private medical and legal chats

The Meta AI app's public Discover feed showed conversations users appeared to have shared by accident, including medical, financial and legal details, because the share flow gave no clear warning. Meta added a warning.

MediumAttack on AIStatus: ContainedVerified case file

Open in the interactive map Cite this case

Reported: Jun 12, 2025
Location: United States
Target: Meta AI app users
Severity: Medium
Triggered by: Meta (product design); users sharing unintentionally
Actor type: Accidental / misconfiguration
Reported loss: None reported
CVE: None

What happened

Impact: Unknown number of users exposed private conversations and identities

Who triggered it

Actor: Meta (product design); users sharing unintentionally

Attribution: Documented by journalists and the Mozilla Foundation; Meta said the feed was not a bug

Motive: None (feature design)

How it happened

Attack vector: Share button that posted chats to a public feed without explicit visibility cues

Root cause: Interface lacked clear indication that shared content was public

Timeline

Jun 9, 2025Reports of sensitive chats on the Discover feed
Jun 2025Mozilla launches campaign with over 10,000 signatures
Jul 3, 2025Mozilla reports Meta added a mandatory warning screen before sharing

Response

Meta added a warning step before sharing; feed remained

Current status: Contained

Sources (open in a new tab)

OECD AI Incidents Monitor oecd.ai
Mozilla Foundation mozillafoundation.org

Cite this case

Meta AI app's Discover feed broadcasts users' private medical and legal chats. AI Threat Watch, case meta-ai-discover-feed, Jun 12, 2025. https://aithreat.io/cases/meta-ai-discover-feed/

Permanent link: https://aithreat.io/cases/meta-ai-discover-feed/