The ChatGPT app includes a new mode designed for those who want to enhance their privacy and security
ChatGPT is one of the most popular AI-powered chatbots. It was among the first to emerge and has since experienced steady growth in its user base. Recently, it added a new feature specifically designed for users who want to enhance their security and prevent data leaks, called "High-Risk Users." Let's discuss the details of this feature.
This new feature restricts some functions in exchange for greater security control. It's not intended for most users who use ChatGPT daily, but rather for those who want maximum protection. For example, people who face certain digital risks, such as journalists, activists, or anyone working in a sensitive environment.
- A new ban has been implemented in ChatGPT
OpenAI has announced a new ban on ChatGPT. In a blog post dated February 13, the company explained how this measure works and its purpose. It clarified that this measure aims to adapt to the rapid advancements in AI systems, which are handling increasingly complex tasks involving websites and connected applications, potentially posing certain risks.
Blocking is an optional advanced security setting for users, providing a level of isolation and protection against data leaks via URLs. However, when this feature is enabled, ChatGPT will not function as usual and will be subject to certain limitations. This feature specifically mitigates point injection attacks, which are considered a serious security vulnerability.
Among these restrictions is that web browsing will rely solely on cached content and will not retrieve real-time information from the internet. The idea behind this restriction is to reduce the risk of unintentional data disclosure while acknowledging that outdated or potentially unintended information may still be obtained.
It should also be noted that the deep search feature will be disabled. Proxy mode, used to perform more independent actions, will also be disabled for users who have enabled it. This is to prevent data leaks and reduce risks.
Other limitations relate to image management; ChatGPT does not include images in its replies, although users can upload them. Additionally, there are restrictions preventing the AI from downloading files for analysis.
- Not a necessary feature for most users
This new feature is available in ChatGPT to users of the ChatGPT Enterprise, ChatGPT Edu, ChatGPT Healthcare, and ChatGPT Teachers packages. In other words, it is not available to users of the basic version of ChatGPT, as it is not designed to meet the needs of the majority of users.
However, OpenAI has stated that this feature will be rolled out to regular user plans in the coming months. So, if you're interested in this new feature, you'll have to wait a few more weeks for it to be activated on your ChatGPT account.
Does this mean that a regular ChatGPT account is insecure because this new feature cannot be enabled, at least for now? In reality, there's no indication of a genuine risk, but data breaches are always a possibility. We've seen how artificial intelligence can impact privacy. Even a vulnerability in Microsoft Copilot exposed sensitive emails .