NewsBite analysis · AI, Security
ZDNet
How ChatGPT's new Lockdown mode protects you from data theft (and what else it does)
OpenAI has rolled out a new Lockdown mode for ChatGPT, designed to protect users from data theft caused by prompt injection attacks. This mode restricts live web access and limits outbound network requests, preventing sensitive information from being exposed during interactions. While it effectively curbs live searches and image retrieval, it does not eliminate the risk of attacks entirely, as malicious commands can still infiltrate cached content. Users are advised to consider this option if handling sensitive data, while being mindful of its operational limitations.
ExperimentConfidence: HighShort term horizon
Published 8 Jun 2026, 16:00 UTC2 min read
Visit original article ↗Lead image — ZDNet
Summary
TL;DR
- Lockdown mode now available for all ChatGPT users.
- Limits live web browsing and outbound network requests.
- Not a complete solution against prompt injection attacks.
- Users can enable Lockdown mode through their account settings.
- Offers protection for sensitive or confidential data interactions.
Interesting facts
- Lockdown mode is a response to rising security threats like prompt injection.
- The mode could limit user experience due to its restrictions.
- Enabling Lockdown mode involves a simple settings adjustment.