2026-11-09 · Cyber Security News · MEDIUM
Google Gemini Allows Download of Filtered AI Images
60-WORD SECURITY BRIEF
Google Gemini’s safety filter bypass lets users download AI‑generated images that the platform’s filters rejected, bypassing controls without a jailbreak. The flaw allows a rejected image to be served while the chat shows an approved one, CVE-2026-19868. Google patched the safety filter logic; users should apply the latest Gemini update or disable the download feature until the fix is deployed.