OpenAI has disrupted scam networks likely originating in Cambodia, Myanmar, and Nigeria that were abusing its models, including ChatGPT, to defraud people online. These operations range from individual actors to organized crime groups, employing a common "ping, zing, sting" tactic: cold outreach, generating enthusiasm or panic, and extracting money or information.

AI is being used as a scaling and efficiency tool by scammers for translation, message writing, and content creation. OpenAI has banned operations posing as investment firms, using AI to create fake websites, social media accounts, and promotional content. Some scammers even asked for advice on masking their location and evading platform restrictions.

Interestingly, OpenAI also observed millions of instances monthly where people use ChatGPT to identify and avoid online scams, estimating this occurs up to three times more often than its use in scams. The company is committed to identifying and preventing such abuses, sharing insights to combat evolving threat patterns.