Last Updated:
OpenAI is offering better safety for ChatGPT users with two new features that will look to stop prompt injections and other attacks.
These two safety features will help users protect their data.
Most users are aware about lockdown modes on their smartphones but with the amount of AI in the open, OpenAI has decided to bring its own version for ChatGPT users. The lockdown mode for ChatGPT guarantees better safety for your data/chats on the AI platform but also alerting them about possible risks where their data can be exposed or exploited.
OpenAI realises the perils of using these AI chatbots, especially with AI agents at work for different tasks. And for that, the company is bringing another feature called Elevated risk labels where any dangers of prompt injections can be handled before they wreak havoc.
Lockdown Mode And Elevated Risk Labels In ChatGPT: All Details
ChatGPT interacts with the internet where OpenAI cannot control how your data is accessed, and the lockdown mode will help the users restrict the reach of their content on external sources. This feature will also block the use of certain tools that need access to the internet as well as talk to other servers for sharing or receiving data.
“This restriction is designed to prevent sensitive data from being exfiltrated to an attacker through browsing. Some features are disabled entirely when we can’t provide strong deterministic guarantees of data safety,” OpenAI mentions in its post.
Having the feature doesn’t mean it will be useful for everyone, and the company points out that advanced security is mostly suited to people who handle sensitive information and data through the AI chatbot.
Lockdown mode is available for ChatGPT Enterprise, ChatGPT Edu, ChatGPT for Healthcare, and ChatGPT for Teachers which can be enabled by IT admins from the workspace settings.
Elevated labels risk as the name suggests is there to alert the ChatGPT users about possible attempts of the AI going rogue.
This feature has been designed to tackle the prompt injection attacks, wherein the hackers can infect prompts to make the AI agents and models deliver malicious files into the targeted system. This is more relevant for its products like ChatGPT Atlas AI browser and Codex that platforms a wide array of AI agents for enterprise tasks.
California, USA
February 17, 2026, 12:36 IST
Read More
