Operation “False Witness”: Fake recovery service impersonating authorities

This case study was originally published in OpenAI’s February 2026⁠(opens in a new window) report on disrupting malicious uses of AI.

Actor

We banned a cluster of ChatGPT accounts using our models to pose as fictitious law firms, as well as impersonate real attorneys and U.S. law enforcement, in a recovery scam targeting fraud victims. This activity very likely originated in Cambodia and aligns with recent public reporting on Chinese-led criminal scam operations in the country.

Behavior

The accounts used ChatGPT to support a fraudulent scam losses recovery operation built around fake law firms and the impersonation of trusted entities, such as real attorneys and the FBI’s Internet Crime Complaint Center (IC3.)

1. Ping: The scammers used ChatGPT to create content that purported to come from at least six fake law firms. Some of this content was then posted by social media accounts and online ads that promoted fictitious scam recovery services. Multiple open-source indicators suggested the firms were fraudulent, including no evidence of state bar licensing, the use of incongruous web domains, contact information at street addresses that do not appear to exist, and directions to contact lawyers via messaging apps.

The scammers used our models to generate social media content promoting fake scam recovery services.

2. Zing: The scammers attempted to gain their targets’ trust by generating comments that tried to emulate the tone and professionalism of attorneys specialized in helping scam victims recover financial losses. They generated content that purported to come from lawyer personas. We identified some of these personas on a range of websites, where their profile pictures were apparently lifted from social media or generated with AI. In some cases, the scammer reused the same “attorney” identities across multiple supposed law firms. One of the websites that we identified posed as IC3.