34 million chat lines. That's the scope of the OmniGPT breach.

The AI chatbot platform — used by over 30,000 users — was compromised in a breach that exposed the complete chat history of its entire user base. Every conversation. Every query. Every piece of information users had shared with the AI, trusting it to remain private.

The exposed data included not just chat content but personal information, email addresses, and metadata that could be used to identify users and reconstruct their AI interactions in detail. For users who had used OmniGPT for sensitive queries — code review, business strategy, personal advice — the exposure was total.

The breach, documented by Skyhigh Security, illustrated the concentrated risk of AI chat platforms: users pour sensitive information into these systems because the interface encourages candid interaction. When that platform is breached, the exposure isn't like a traditional data breach — it's a complete record of everything the user was thinking and asking about.

34 million lines of conversations that users thought were private. They weren't.