According to monitoring by Beating, OpenAI has announced the expansion of ChatGPT’s Lockdown Mode—previously available only to enterprise and education users—to all individual and commercial accounts. As a final line of defense against prompt injection attacks, Lockdown Mode is now available across Free, Plus, Pro, and ChatGPT Business tiers. Under Lockdown Mode, several online and analytical features are disabled: real-time web browsing is restricted to cached content only, Deep Research and Agent Mode are fully disabled, Canvas-generated code is prevented from accessing the internet, and file download channels used for data analysis are disconnected. Additionally, functions for retrieving and displaying images from web pages are restricted, though users can still generate and upload images normally. Prompt injection attacks rely on embedding malicious instructions within web content or uploaded files to trick large models into performing unintended actions. Security expert Simon Willison noted that Lockdown Mode severs the data exfiltration channel among the “fatal trio” of attack vectors. Unlike traditional defenses that depend on the AI’s own risk assessment, Lockdown Mode establishes a deterministic protective framework by physically blocking network requests, rendering it immune to malicious prompts. However, OpenAI emphasizes that Lockdown Mode does not prevent prompt injection from affecting the model’s logical behavior or response accuracy—it only mitigates the risk of sensitive data leakage. Lockdown Mode was initially announced in February 2026 and was originally limited to enterprise and education subscribers. OpenAI’s Chief Information Security Officer, Dane Stuckey, stated that Lockdown Mode is not intended for all users but is a security tool specifically designed for high-risk groups. While it involves trade-offs in functionality and usability, it remains a valuable measure for these targeted users.
OpenAI Expands ChatGPT Lockdown Mode to All Users, Disables Certain Online Features
MarsBitShare
Summary
OpenAI has rolled out ChatGPT Lockdown Mode to all users, including personal and commercial accounts. This mode disables features such as real-time web browsing, Deep Research, and Agent Mode to prevent data exfiltration. News related to real-world assets (RWA) may now be handled with increased caution under this setting. The update, first announced in February 2026, is now available across all ChatGPT versions. OpenAI states that the mode does not block malicious prompts but reduces exposure of sensitive data. Queries involving inflation data and other online research functions are affected by this change.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.