OpenAI Introduces Lockdown Mode for ChatGPT to Protect Against Prompt Injection Attacks

2 source(s) · earliest article 2026-06-06 18:00 +04:00 · cluster #1194 · category: Tech · model: qwen2.5:7b · priority 44.8/100

OpenAI has introduced Lockdown Mode as an additional security feature for ChatGPT to protect against prompt injection attacks. The mode disables live web browsing, image retrieval from the web, deep research, and agent mode. While it aims to reduce data exfiltration risks, OpenAI notes that ChatGPT could still be vulnerable in certain scenarios. Meanwhile, Meta’s AI app now features a For You section with AI-generated clickbait-style stories, which Meta has since removed following criticism.

What each source uniquely contributed

• The Verge lean-left · tech — OpenAI announced Lockdown Mode to protect against prompt injection attacks and detailed its limitations.
• TechCrunch lean-left · tech — Meta’s AI app now includes an AI-generated For You section with clickbait-style stories, which Meta has since removed.

Framing differences

• OpenAI focused on security and data protection, while The Verge highlighted the removal of a controversial feature by Meta.

All sources

1. [S1] The Verge: Meta made its own AI-generated clickbait news feed lean-left · tech · 2026-06-06 18:00 +04:00
2. [S2] TechCrunch: OpenAI unveils Lockdown Mode to protect sensitive data from prompt injection attacks lean-left · tech · 2026-06-07 00:32 +04:00