OpenAI Disrupts Cambodia Scam Network Using ChatGPT
SkimNews Take
AI compresses the operational cost of running diverse fraud verticals — investment, romance, gambling, impersonation — into a single toolkit, letting one coordinated group scale the variety of scams that previously required separate specialized teams.
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI banned a coordinated cluster of ChatGPT accounts operating from Poipet, Cambodia, using its models to fabricate personas, translate scam messages, and generate promotional content for fraudulent operations
- The network ran investment, romance, gambling, and law enforcement impersonation schemes in parallel, blending techniques across categories — including posing as police to pressure targets into paying fictitious fines
- Some accounts posted social media ads recruiting "chatter" workers in Bangladesh and India, offering an $800 base salary plus a $100 "full attendance" bonus, with flights, Cambodia visas, accommodation, and meals included
- A subset of accounts handled administrative back-office work — drafting internal announcements, translating between staff, and documenting employee debts, salary deductions, fines, loan repayments, and visa overstays — patterns OpenAI says are consistent with human trafficking and forced labor in East Asian scam compounds
- The scammers followed a three-step "ping-zing-sting" attack chain: initial outreach on WhatsApp or Telegram, a trust-building phase, then instructions to deposit funds or settle non-existent fines, followed by fake transfer screenshots as proof of payment
- OpenAI investigated the operation in partnership with Meta-owned WhatsApp; the company said financial losses are unverified but internal communications referenced individual victims losing thousands of dollars across multiple scam types
Why it matters: The discovery that ChatGPT was used not just to con victims but to run the internal HR of a trafficking compound — tracking worker debts, fines, and visa status — shows generative AI is now embedded across the full stack of industrial-scale scam operations, from recruitment ads to back-office administration. OpenAI's takedown with WhatsApp signals platforms are starting to coordinate against AI-augmented fraud networks that previously operated with little visibility.
Ask SkimNews
