Android Notification Vulnerability Allowed Malicious Prompts to Hijack Google Gemini
SafeBreach researcher Or Yair discovered a vulnerability called "Fake Context Alignment" where a single malicious notification on Android could trick Google Gemini into performing harmful actions like faking messages, opening windows, joining Zoom calls, and poisoning long-term memory — all without requiring any malicious app to be installed. Google has since patched the issue with server-side mitigations after the researcher reported it through Google's Vulnerability Reward Program.
Key quotes
SafeBreach researcher Or Yair showed that a single poisoned notification on Android could trick Google Gemini into faking messages, opening windows, joining Zoom calls, and even poisoning long-term memory without any malicious app installed.
Google has since patched the issue with server-side mitigations after the Fake Context Alignment bypass was reported through its Vulnerability Reward Program.
Gemini on Android could treat hostile notifications as instructions.
You might also wanna read
Security Vulnerability: Google's Antigravity AI Susceptible to Indirect Prompt Injection Attacks
The article describes a security vulnerability where Google's Antigravity AI system (likely referring to Gemini) can be manipulated through
promptarmor.com·7mo agoGoogle Gemini is about to control your messages and calls, even if you say no
Google expands testing of persistent Gemini overlay bubble with refreshed gradient design on Android
Google is expanding testing of a persistent Gemini overlay bubble in Android, which allows users to keep their Gemini conversation active ev
Google Gemini Adds AI-Generated Image Detection Feature
Google is introducing a new feature in its Gemini app that allows users to detect AI-generated images, but only those created or edited by G
The Verge·7mo ago
Google Chrome Installs 4GB Gemini Nano AI Model on Users' Computers Without Consent
Google Chrome has been secretly installing Gemini Nano, a 4GB on-device AI model, onto users' computers without their knowledge or consent.
oztalking.com·3d ago
Google's Gemini AI push risks repeating Microsoft's Copilot mistakes
The article criticizes Google's aggressive push of Gemini AI features across its product ecosystem, drawing parallels to Microsoft's widely
The Verge·1mo ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.