OpenAI Shuts Down AI-Powered Scam Network
OpenAI has shut down a network of accounts that used ChatGPT to help run a large-scale online scam operation based in Southeast Asia. The group, believed to be operating from the Cambodian city of Poipet, used artificial intelligence to create fake identities, write convincing messages, translate conversations, and manage daily operations for a variety of online scams.
According to OpenAI, the scammers relied on AI to make their fraud schemes appear more believable and efficient. They created fake social media profiles, wrote messages designed to gain victims’ trust, and produced promotional content for fraudulent investment opportunities, online gambling offers, and romance scams. The network also used AI to create fake documents, including passports, legal notices, and financial records, to make their scams seem legitimate.
The group reportedly targeted people through popular messaging apps such as WhatsApp and Telegram. Scammers would often begin with friendly conversations or romantic relationships to build trust before encouraging victims to invest in fake cryptocurrency or gold trading platforms. Others pretended to represent online gambling services, promising fake bonuses or winnings, while some impersonated law enforcement officers and demanded payments by falsely claiming victims owed fines or had committed crimes.
Investigators also found that the scammers used ChatGPT for administrative tasks within the organization. AI helped draft internal announcements, translate conversations between employees, track salaries, record debts and fines, and manage recruitment efforts. Some of the recruitment advertisements promised high-paying jobs in Cambodia with free housing, travel, and work permits, but researchers believe these offers may have been connected to human trafficking operations that force workers into carrying out online scams.
OpenAI said it worked with WhatsApp to investigate the operation and remove the accounts involved. While the total financial losses are unknown, conversations recovered during the investigation suggest that the network may have targeted hundreds of victims, with some individuals reportedly losing thousands of dollars. These claims have not been independently verified.
The investigation highlights how cybercriminals are increasingly using artificial intelligence to make scams faster, more convincing, and easier to scale. By automating conversations, creating realistic fake identities, and generating convincing documents, AI allows scammers to reach more potential victims with less effort. As AI technology continues to evolve, experts say organizations and individuals should remain cautious of unexpected messages, investment opportunities that seem too good to be true, and requests for personal or financial information from unfamiliar sources.
