AI Threat Watch

Case omnigpt-alleged-breachModel & data leakWorldwide

Hacker 'Gloomer' claims OmniGPT breach leaking 34M chat lines and 30K user contacts

A BreachForums user posted data allegedly taken from OmniGPT: about 34 million lines of user-chatbot conversations, emails and phone numbers of roughly 30,000 users and links to uploaded files containing API keys and credentials.

HighAttack on AIStatus: UnresolvedVerified case file

Open in the interactive mapCite this case

Reported
Feb 9, 2025
Location
Worldwide
Target
OmniGPT (multi-model AI chat aggregator)
Severity
High
Triggered by
Gloomer (BreachForums persona)
Actor type
Criminal
Reported loss
None reported
CVE
None

What happened

A BreachForums user posted data allegedly taken from OmniGPT: about 34 million lines of user-chatbot conversations, emails and phone numbers of roughly 30,000 users and links to uploaded files containing API keys and credentials.

Impact: If authentic, 34,270,455 lines of conversations plus contact details for about 30,000 users and uploaded files holding API keys, credentials and billing data were exposed

Who triggered it

Actor: Gloomer (BreachForums persona)

Attribution: Self-claimed by the forum user Gloomer; OmniGPT did not confirm or deny the breach and no independent verification was published

Motive: Sale and leak of user data for profit and notoriety

How it happened

Attack vector: Unknown; the actor did not disclose how OmniGPT's backend was accessed

Root cause: Unknown; researchers noted that, if genuine, the dump would point to weak authentication and data protection on the platform

Timeline

  1. KrakenLabs flags an actor offering OmniGPT data for sale for about 100 dollars
  2. Dataset posted on BreachForums under the Gloomer alias
  3. CyberInsider and SiliconANGLE report the claims; OmniGPT makes no public statement

Response

No public statement from OmniGPT at the time of reporting; outlets advised users to reset passwords, revoke API keys and enable MFA

Current status: Unresolved

Sources (open in a new tab)

Cite this case

Hacker 'Gloomer' claims OmniGPT breach leaking 34M chat lines and 30K user contacts. AI Threat Watch, case omnigpt-alleged-breach, Feb 9, 2025. https://aithreat.io/cases/omnigpt-alleged-breach/

Permanent link: https://aithreat.io/cases/omnigpt-alleged-breach/

This case file in other languages: EnglishFrançaisEspañolالعربية中文