Google Gemini Safety Filter Bypass Allows Download of Blocked Images
First merged summary.
Google Gemini safety filter bypass exposed blocked AI images via download links, resolved by maintenance without exploitation.
Fog Security researchers found the Google Gemini web app permits downloading AI-generated images rejected by safety filters through separate download paths. After a refused image in the chat interface, the download returned the blocked artifact while the preview showed an approved version. The issue was reported to Google in June 2026 but could no longer be reproduced by October 22, 2026 following routine maintenance. No privilege escalation, cross-user exposure, or data access occurred.
- Separate paths for image viewing and download
- Blocked images retrievable via downloads
- Reported to Google in June 2026
- Resolved by October 22, 2026 via maintenance
- No jailbreak or privilege escalation required
- Tested on copyrighted characters and synthetic images
Timeline
dated events in the reporting · oldest first- Oct 22, 2026Issue no longer reproducible after maintenance · Cyber Security News
- Nov 9, 2026Reports published on Gemini download link bypass · GBHackers and Cyber Security News
What each outlet adds
beyond the shared facts- GBHackersSeparate paths for viewing and download; behavior changed after maintenance
- Cyber Security NewsTested on copyrighted characters and own likeness; resolved by October 2026
- Vendors
- Products
- Gemini
- Organizations
- Fog Security
- AI models
- Gemini
Coverage
oldest first · one row per article · UTC- · 4h ago
GBHackersFirst report
Google Gemini Safety Filter Bypass Allows Blocked Image Downloads
Gemini exposed rejected images via download links after safety filter refusal.
- · 3h ago
Cyber Security News
Gemini allows download of blocked AI images
Gemini safety filter sometimes delivered refused images via download path after chat showed permitted preview; issue resolved via maintenance.