New ChatGPT Lockdown Mode Limits Tools That Could Enable Data Exfiltration

Summary: OpenAI has begun rolling out an optional Lockdown Mode for ChatGPT on eligible personal accounts and self-serve ChatGPT Business plans, aimed at reducing data exfiltration from prompt injection attacks by limiting outbound network requests. The setting disables or limits tools that connect to the web or external services, cannot be used at the same time as Developer Mode, and OpenAI states it does not guarantee exfiltration cannot happen, with risk remaining through enabled Apps, unforeseen combinations of capabilities, or newly discovered techniques.
What you should know: OpenAI's own caveat is the stake: Lockdown Mode substantially reduces but does not guarantee prevention, and a malicious instruction hidden in an uploaded file can still alter ChatGPT's behavior, so the residual risk sits with the organization enabling it.

Read the original at The Hacker News ↗

Topics: AI Security
Related: SECURITY AFFAIRS AI-CYBERSECURITY NEWSLETTER ROUND 8 Cyber Risks · Nov 15, 2026
Pentagon Renames CDAO ‘Super Intelligence Office’ Cyber Risks · Nov 15, 2026
186 Fake AI Domains Redirect Users to Click Fix Lures, Browser Malware, and Search Hijackers Cyber Risks · Nov 12, 2026