ChatGPT has a new security feature called Lockdown Mode, but OpenAI is clear about one thing from the start: most people don’t need to turn it on.
The new mode is designed for users who face unusually high levels of digital risk, such as journalists, activists or people working in sensitive environments. According to OpenAI, the standard protection measures integrated into ChatGPT are already sufficient for everyone else.
What is lock mode in ChatGPT?
Lockdown mode is essentially a high security setting that limits what ChatGPT can do. When enabled, the system behaves more conservatively to reduce the risk of data exploitation, manipulation, or inadvertent disclosure. OpenAI describes this as reducing the product’s attack surface, even if it means limiting useful features.
One of the biggest changes concerns web access. Internet browsing is limited to cached content only, which prevents ChatGPT from retrieving live information from the Internet. This reduces the risk of sensitive data being transmitted, but also means that search results may be incomplete or out of date.
Image processing has also been tightened. ChatGPT can no longer include images in its replies, but users can still upload their own images and continue to use image generation tools.
Deep Research, which enables more complex multi-level analysis, is completely disabled, as is Agent Mode, which would otherwise allow the system to perform more autonomous actions.
Other restrictions focus on networks and files. Canvas-generated code cannot be allowed to access the network and ChatGPT cannot download files for data analysis. It can still work with files you upload manually, but it doesn’t retrieve anything on its own.
Taken together, these limits are intended to reduce the system’s external reach or its independent action.
Who can currently use lockdown mode?
Lockdown mode is currently available for ChatGPT Enterprise, Edu, ChatGPT for Healthcare and ChatGPT for Teachers. OpenAI plans to roll out the feature to Consumer and Team plans in the coming months.
For organizations, lockdown mode is managed at the workspace level. Administrators can create a custom role specifically for lockdown mode and assign it to select users, allowing for tighter controls without changing settings for everyone.
In other ChatGPT news, the platform recently added the ability to drive deep research using content from your own websites and apps, giving more control over how results are collected. At the same time, changes to ad settings for free accounts mean you can avoid ads, but your usage limits may shift.
