A recently disclosed vulnerability in Google’s Gemini AI panel could have allowed hackers to hijack the feature and access sensitive data on a user’s device. Researchers at Palo Alto Networks’ Unit 42 first discovered the flaw, which is labeled as CVE-2026-0628.
According to the report, the issue stemmed from how Chrome handled permissions for the Gemini side panel. This is a browser feature that integrates Google’s AI assistant directly into the browsing experience. The discovered vulnerability could have enabled malicious browser extensions with basic permissions to inject code into the Gemini panel.
Since the Gemini panel runs with elevated privileges in Chrome, attackers could exploit the flaw and gain access to systems that are normally restricted.
What hackers could’ve done with the exploit
Once the Gemini panel is hijacked, the attacker can potentially execute code with powerful system-level privileges. Researchers showed that this would enable several dangerous actions, such as:
- Accessing the camera and microphone without user consent
- Taking screenshots of any webpage
- Reading local files and directories from the OS
- Running malicious scripts inside the Gemini interface
The good news: Google already patched it
The vulnerability was initially disclosed to Google in October 2025, and the company released a fix in January 2026 after reproducing the issue internally. While the flaw is now patched, security researchers warn that the incident highlights a broader issue, which is that AI-powered browser features introduce new security risks because they require deeper access to the system.
So for the everyday user, the takeaway is simple. Update Chrome immediately to make sure you’re on a version that includes the security fix.