OpenAI introduces Lockdown mode in ChatGPT to block prompt injection
OpenAI just rolled out Lockdown Mode for ChatGPT, designed to protect users from prompt injection attacks, basically, sneaky tricks that can make AI do things it shouldn't or leak private information.
This new mode is especially handy if you're working with sensitive stuff and want an extra layer of security beyond the usual.
ChatGPT lockdown mode pauses key features
Turn on Lockdown Mode, and some features get paused: no fetching internet images, no file downloads for analysis, and advanced tools like Deep Research and Agent Mode are switched off.
The essentials, like uploading files and chat memory, still work.
Anyone using ChatGPT, even free accounts, can try it out in Safety settings. Plus, there's a new session manager so you can see where you're signed in and log out remotely if needed.