News · curated 8 Sep 2026

Extortion crews have their eyes on high-value AI data, Google warns

Coverage timeline

8 Sep 2026theregister.com

Single-source analysis — first reported, latest, and curated coincide.

Why it matters

Google's warning signals that AI models, training data, and intellectual property are now high-value extortion targets, raising the stakes for organizations that concentrate valuable AI IP in breachable systems.

Google's Mandiant threat hunters, in their latest AI Threat Tracker, report that data-theft-and-extortion crews are increasingly stealing companies' proprietary AI data — including source code, prompts, model scripts, secrets, and proprietary models — and threatening to leak it unless a ransom is paid. Cited cases include a healthcare firm whose drug and AI research plus a proprietary model were exfiltrated, and an AI media-generation company whose AI assets were stolen, with actors like TeamPCP named as especially successful.