ChatGPT Data-Leak Flaw Exposed User Conversations and Sensitive Information

ChatGPT is widely used for text generation, document analysis and workplace collaboration, and users may share personal data, business documents and other confidential material in chats or uploaded files. Cybersecurity company Check Point said attackers who bypass the platform's safeguards could quietly send data to external systems without users' knowledge, highlighting the security risks of processing sensitive information with generative AI.

Check Point researchers recently disclosed a ChatGPT data-leak vulnerability that allowed attackers to use DNS tunneling to circumvent existing security measures and steal users' conversation histories and sensitive uploaded data. OpenAI patched the flaw on February 20. Although the issue has been addressed, the researchers warned users to avoid entering personal information, login credentials or nonpublic corporate information.