ChatGPT Lockdown Mode: Prevents Prompt Injection Attacks

ChatGPT has rolled out a new privacy feature called Lockdown Mode to all users, including those on the free tier. This mode helps prevent prompt injection attacks, which are a growing concern in the AI sector. Lockdown Mode stops ChatGPT from making live outbound network requests. This prevents malicious actors from tricking the AI into sending personal information to third parties. The feature was first introduced in February for enterprise users. It is now available for Free, Plus, Pro, Go, and self-serve Business accounts. Turning on Lockdown Mode does mean some ChatGPT features that rely on external calls will be restricted. However, for users handling sensitive data, the extra security is often a worthwhile trade-off. This is important because prompt injection attacks exploit AI instructions, not software vulnerabilities, by placing malicious commands within text. This update allows users to choose greater control over their data and privacy.