
OpenAI said the banned accounts were tied to a Poipet-based criminal network that used ChatGPT to build fake identities, produce scam materials and support bogus crypto and gold investment schemes.
OpenAI said it terminated ChatGPT accounts linked to a criminal network in Poipet, Cambodia, that used the chatbot to create fake identities, translate conversations, draft trust-building messages and produce marketing materials for bogus cryptocurrency and spot-gold investments. The operation, identified after a tip from WhatsApp, allegedly steered victims from romance or friendship conversations to fake trading platforms, then extracted additional payments through supposed activation fees, taxes and regulatory fines. OpenAI also said the investigation uncovered signs of labor trafficking, including recruitment ads for “chatters” and spreadsheets tracking workers’ debts, salary deductions, fines, immigration status and attempted escape. The case adds to concerns that organized scam networks are combining generative AI with crypto payment rails to scale fraud, even as researchers and blockchain analytics firms warn that Southeast Asian pig-butchering operations increasingly rely on both trafficked labor and digital assets.