Artificial intelligence is increasingly becoming a double-edged sword, and a recent incident involving **OpenAI**'s **ChatGPT** illustrates this stark reality. The company revealed it had shut down accounts linked to a Cambodian cyber scam network that was extensively using its AI models. ### AI at the Core of Deception The scammers employed **ChatGPT** to facilitate various fraudulent schemes. These included investment fraud, romance scams, gambling scams, and even impersonation of law enforcement officials. A particularly egregious aspect was the creation of fake job advertisements targeting individuals in India, promising enticing benefits like free flights and visas, ultimately luring victims into forced labor scenarios. **OpenAI** stated in a [blog post](https://openai.com/index/disrupting-malicious-uses-of-ai-criminal-scam-operation/) that their investigations frequently observe malicious actors adapting between scam types or combining multiple techniques within a single operation. > "The network used our models to create and support the operation of fake online personas, generate and translate messages sent to scam targets, create promotional content for their fraudulent schemes, and assist with day-to-day operations." ### A Hub of Illicit Activity The activity was traced back to Poipet, a city in Cambodia known for its deep ties to scam operations often run by Chinese criminal gangs. Beyond direct victim interaction, **ChatGPT** was also used to generate internal communications for the scam center, including staff announcements, translation between members, and recruitment flyers. Furthermore, the AI was used to create fake documents such as passport images, legal notices, stock purchase confirmations, and gambling platform pages, adding a layer of authenticity to the fraudulent schemes. ### The Human Cost of AI-Powered Scams The scope of the financial losses remains unquantified by **OpenAI**, but the operation is believed to have engaged with hundreds of targets. Disturbingly, internal records maintained by the scammers, translated with **ChatGPT**, referenced employee debts, salary deductions, disciplinary fines, and loan repayments. Conversations also alluded to detention, escape attempts, and potential criminal liability for those who had been trafficked and forced into these operations. ### A Widespread and Growing Threat This incident underscores a broader trend of proliferating scam centers across Southeast Asia, despite efforts from U.S. and Chinese law enforcement to disrupt them. Experts are now warning that these operations are expanding into South Asia and Africa. Reinforcing this concern, **INTERPOL** recently published a study indicating that over half of all cybercrime in Africa now incorporates artificial intelligence. The agency surveyed 36 member states, with law enforcement believing AI is "now used to automate every stage of the attack lifecycle: from reconnaissance and phishing to extortion and evasion." One particularly alarming case in Uganda involved AI-generated audio and video being used to impersonate a public figure, leading to over $2 million in losses through a fraudulent investment scheme.