Case omnigpt-breach-claimModel & data leakUnited States
Hacker claims OmniGPT breach exposing 34 million chat messages and 30,000 user emails
A BreachForums user named Gloomer posted samples allegedly taken from OmniGPT: over 34 million lines of user-chatbot conversations, around 30,000 emails with some phone numbers, file links and credentials. Unconfirmed.
HighAttack on AIStatus: UnresolvedVerified case file
Open in the interactive mapCite this case
- Reported
- Feb 9, 2025
- Location
- United States
- Target
- OmniGPT AI chatbot aggregator
- Severity
- High
- Triggered by
- Gloomer (BreachForums poster)
- Actor type
- Criminal
- Reported loss
- None reported
- CVE
- None
What happened
A BreachForums user named Gloomer posted samples allegedly taken from OmniGPT: over 34 million lines of user-chatbot conversations, around 30,000 emails with some phone numbers, file links and credentials. Unconfirmed.
Impact: Alleged exposure of millions of private AI conversations, uploaded documents, API keys and crypto keys
Who triggered it
Actor: Gloomer (BreachForums poster)
Attribution: Self-claimed by the poster; samples reviewed by researchers but not verified by the vendor
Motive: Data sale and notoriety
How it happened
Attack vector: Unknown; attacker claimed access to backend message store and user tables
Root cause: Unknown
Timeline
- Data samples posted on BreachForums
- Wide press coverage; OmniGPT silent
Response
No public statement from OmniGPT found
Current status: Unresolved
Sources (open in a new tab)
- CyberInsider cyberinsider.com
- SiliconANGLE siliconangle.com
- CSO Online csoonline.com
Cite this case
Hacker claims OmniGPT breach exposing 34 million chat messages and 30,000 user emails. AI Threat Watch, case omnigpt-breach-claim, Feb 9, 2025. https://aithreat.io/cases/omnigpt-breach-claim/
Permanent link: https://aithreat.io/cases/omnigpt-breach-claim/