Google Patches Android Lock Screen Security Flaw In Gemini AI

A critical security flaw in Android 16 allowed unauthorized individuals with physical device access to bypass PIN authentication via Google's Gemini AI. By using specific multi-touch gestures, such as pressing a "Continue" prompt alongside an attachment button on the lock screen, users could interact with sensitive applications. This allowed attackers to send SMS and WhatsApp messages or extract personal data without entering a passcode. The bug was confirmed to affect various devices, including the Pixel 6a, though reports suggest impact varied across manufacturers.

The vulnerability originated from a software-level oversight in how Gemini manages app-handoffs when the assistant is active on the lock screen. Analysts noted that executing specific timing commands allowed the AI to break security boundaries, effectively granting full application functionality from a locked state. While some users reported difficulty reproducing the bug on specific hardware like Samsung phones, the flaw posed a major risk for phone theft and potential extortion, as it enabled malicious actions before the device could be recovered.

Google confirmed the vulnerability and initiated a mandatory security update to close the loophole across affected Android models. Until the patch is fully installed, cybersecurity experts advise users to disable Gemini’s lock screen permissions in the phone settings. Users should toggle off the options for using Gemini and sending messages without unlocking the device. Moving forward, this incident places renewed pressure on Google to harden its security architecture as it integrates generative AI more deeply into the Android ecosystem.