Check Point Research finds bypass of ChatGPT Lockdown Mode

Check Point Research has disclosed a vulnerability that bypassed ChatGPT Lockdown Mode, a feature intended to limit prompt-injection-driven data exfiltration. The flaw could have allowed silent theft of chat data, phishing and malware delivery, including zero-click exfiltration in the ChatGPT desktop application.
OpenAI fixed the issue after responsible disclosure. Check Point said the finding points to a wider enterprise problem: AI security controls are only as strong as the application environment around them.