OpenAI banned accounts using AI to support romance scam workflows, including outreach, translation, victim engagement, and investment-fraud lures. Since beginning to report on scam network disruptions, OpenAI has taken down numerous operations worldwide, including "task" scams and investment scams.

Romance scams predate AI, with scammers posing as romantic partners to extract money from targets. OpenAI's first published scam disruption targeted a criminal operation in Cambodia that used ChatGPT to generate romance scam messages spread on social media.

These scams follow a common AI abuse pattern: the ping (cold contact), the zing (generate emotion), and the sting (extract money). The ping uses AI-generated content to attract targets—such as the "pig butchering" scam that targeted American men in their 40s in medical professions via golf-related social media replies. The zing generates emotionally manipulative content. The sting convinces targets to hand over money through fake investments, business crises, or fictional task earnings.

Distribution networks span SMS, encrypted messaging apps, social media posts, and online ads. Operations like "Wrong Number," exposed in June 2025, used cold-call SMS with implausibly high returns to create emotional manipulation.