OpenAI Patches ChatGPT Data Exfiltration Flaw and Codex GitHub Token Vulnerability

Check Point has identified a previously unknown vulnerability in OpenAI's ChatGPT that enables malicious prompts to covertly exfiltrate sensitive user data, including messages and uploaded files, without user awareness. This flaw allows a single malicious input to transform normal conversations into data leakage channels, posing significant privacy and security risks for users of the AI model.