Hundreds of targets may have been approached by a Cambodia-based scam network that OpenAI says it disrupted after finding ChatGPT was used across investment, romance, gambling and fake law-enforcement schemes.
WhatsApp provided the initial lead, and OpenAI said the operators used its models to build fake personas, translate and script messages, forge document images, and handle internal administrative work.
Poipet job ads promising flights, visas and accommodation, along with records of debts, salary deductions, fines and escape-related discussions, pointed to possible forced labor and human-trafficking links inside the operation.
OpenAI banned the accounts, shared threat indicators with industry partners and authorities, and said the case shows scam networks increasingly blend multiple fraud types with broader organized-crime activity.