OpenAI introduces Lockdown Mode to guard against prompt injection attacks
5 reported
OpenAI has announced a new feature called Lockdown Mode, designed to provide additional protection against prompt injection attacks, where malicious instructions are hidden in webpages and other content sources. The feature disables live web browsing, retrieval and display of images from the web, deep research, and agent mode, though users can still generate images. OpenAI stated that even with Lockdown Mode enabled, ChatGPT could remain vulnerable to prompt injections that may appear in cached web content or uploaded files, potentially affecting response behavior or accuracy. The company emphasized that Lockdown Mode is not intended for everyone, but rather for individuals and organizations handling sensitive data who want stricter protection from data exfiltration risks related to prompt injection. OpenAI said it is currently rolling out Lockdown Mode to self-serve ChatGPT Business accounts and eligible personal accounts.
What’s reported
Lockdown Mode disables live web browsing, retrieval and display of images from the web, deep research, and agent mode.
Users can still generate images with Lockdown Mode on.
OpenAI stated that ChatGPT could still be vulnerable to prompt injections in cached web content or uploaded files.
The feature is designed for people and organizations handling sensitive data.
Lockdown Mode is rolling out to self-serve ChatGPT Business accounts and eligible personal accounts.
Sources: TechCrunch
