Mercor-LiteLLM Data Breach
Coverage from All About Cookies, Federman & Sherwood, and others

A malicious update to the open-source LiteLLM tool gave attackers a route into systems associated with Mercor, an AI staffing and data-training company.
Reported exposed material includes contractor identity documents, facial and voice biometrics, interview recordings, Social Security numbers, source code, credentials, and internal data. Mercor commissioned forensic investigations and reported limited exposure, while independent reporting and lawsuits described a larger potential theft and alleged inadequate safeguards. The incident also affected commercial relationships, including reported scrutiny or suspension of work by Meta, and led to class-action litigation in California and Texas.
