OpenAI Disrupts Cambodia-Based Scam Network That Used ChatGPT for Investment, Romance, and Impersonation Schemes
This digest was compiled by AI from multiple sources — links to the originals are below.
OpenAI said it disrupted a coordinated network of ChatGPT accounts operating from Poipet, Cambodia, that carried out investment, romance, gambling, and law enforcement impersonation scams. The accounts were used to create fake online personas, generate and translate scam messages, and produce promotional materials to lure victims globally. The disruption was carried out in partnership with Meta's WhatsApp.
Network Structure
The operation was based in Poipet, a Cambodian city known for scam compounds and human trafficking. OpenAI banned a cluster of accounts likely originating from Southeast Asia that used ChatGPT for a wide range of fraudulent activities. The network created social media ads for "chatter" jobs targeting users in Bangladesh and India, offering $800 monthly salary plus bonuses, flights, accommodation, and work permits. Some accounts also leveraged ChatGPT for internal administrative work such as drafting announcements, translating messages, and documenting employee debts and visa overstays.
Fraud Methods
The group ran multiple scams in parallel, blending tactics to maximize deception. Investment fraud involved building trust through dating personas before directing victims to fake cryptocurrency or spot gold trading platforms. Romance scams used synthetic identities for extended conversations, while gambling schemes promised fake bonuses and winnings. Law enforcement impersonation scams pressured targets into paying fines for fabricated offenses. The network also generated images of forged passports, legal notices, and stock-purchase confirmations.
AI-Assisted Operations
OpenAI's investigation, conducted with Meta's WhatsApp, revealed how the network exploited ChatGPT for operational efficiency. The AI was used to generate and translate scam messages, create social media advertisements, and assist with day-to-day tasks. This illustrates a broader trend of organized crime groups adopting generative AI to scale their deceptive operations.
What's Next
OpenAI said it will continue to refine its detection systems to counter such threats, but the adaptability of transnational scam networks poses an ongoing challenge. It remains unclear how many victims were affected and whether the disrupted actors will reemerge under different guises.
3 sources
OpenAI Disrupts Cambodia-Based Scam Network That Used ChatGPT for Investment, Romance, and Impersonation Schemes
