Researchers Bypass ChatGPT Lockdown Mode to Exfiltrate Chats Without User Interaction

Researchers discovered a patched vulnerability in OpenAI's ChatGPT Lockdown Mode that could bypass protections and enable silent chat data theft and phishing attacks. The flaw has already been fixed, but highlights risks from prompt injection attacks. This is relevant to email security professionals concerned with AI-driven threats.

https://thehackernews.com/2026/11/researchers-bypass-chatgpt-lockdown.html