OpenAI announces Lockdown Mode to reduce prompt injection risks for sensitive data handling

OpenAI rolled out Lockdown Mode to self-serve ChatGPT Business accounts and eligible personal accounts. The feature disables live web browsing, image retrieval from the web, deep research, and agent mode (autonomous task execution).

Lockdown Mode reduces—but does not eliminate—the risk that prompt injections (malicious instructions hidden in webpages or uploaded files) could cause ChatGPT to share sensitive data or alter response accuracy.

OpenAI states the feature is designed for people and organizations handling sensitive data who want stricter protection from data exfiltration risks; the company notes it is not intended for general users.