Scam crackdown
OpenAI has banned a number of ChatGPT accounts it says were tied to bogus law firms and fake lawyers running so-called “scam recovery” schemes.
The activity features in OpenAI’s latest threat update, which outlines an operation it dubbed ‘False Witness’. According to the tech giant, fraudsters used ChatGPT to help draft convincing legal-sounding messages, build credible online profiles and steer victims towards paying fees, often in cryptocurrency.
The scam typically targets an unsuspecting member of the public who has already lost money to fraud and is now searching for legal help. Up pops a professional-looking website offering specialist recovery services, complete with lawyer profiles, polished language and confident promises of action. Then, just as hope starts to creep back in, the conversation shifts to payments.
OpenAI says the network promoted at least six supposed law firms. In some instances, it claims the fraudsters went further, impersonating real lawyers and even law enforcement bodies. Accounts were used to translate messages, rewrite communications into “American English”, produce text “in the style of a lawyer” and even generate fake supporting documents.
The tech giant says it will continue to disrupt malicious uses of its models and ban accounts linked to scams and impersonation.
