OpenAI Introduces Lockdown Mode to Protect Against Cyberattacks

OpenAI has announced a new Lockdown Mode feature for ChatGPT users. This mode is designed to provide additional protection for the chatbot against hidden malicious instructions in web pages and other content sources—known as "prompt injection" attacks. According to Techcrunch.com reports .

When Lockdown Mode is enabled, several system capabilities are restricted. Specifically, live web browsing is disabled, allowing only the use of cached data. Additionally, searching and displaying images from the internet, as well as Deep Research and Agent Mode functions, are temporarily suspended. It is worth noting that users will still be able to generate images.

According to company representatives, even with this mode enabled, ChatGPT may not be completely secure. There is a possibility that malicious commands stored in cached content or uploaded files could affect the accuracy of responses. However, the main goal is to minimize the risk of sensitive data leakage to the greatest extent possible.

OpenAI emphasizes that Lockdown Mode is not intended for all users. It is primarily beneficial for organizations and individuals who handle highly sensitive data and require strict protection against data exfiltration. Currently, this feature is being gradually rolled out for ChatGPT Business and select personal accounts.