Chrome Canary Tests Gemini Password Change Automation

Chrome already supports automated password changes on some websites after it detects a compromised password. Android Authority characterizes the new Canary work as an extension of that workflow to weak and reused credentials. The publication also notes that code references to AI-assisted password tools had appeared in Chromium Gerrit commits in 2024.

Google announced automatic password changes for weak or compromised credentials at Google I/O 2025, The Verge reported. Under that announced workflow, Google Password Manager can prompt a user to fix a detected credential and, on supported websites, generate and update a replacement password. Parisa Tabriz, Chrome's VP and GM, told The Verge that Chrome would not change a bad or compromised password without user consent.

The experiment matters because password managers can identify risky credentials, but correcting them typically requires a site-specific sequence of authentication, password-policy validation, form submission, and confirmation. An automated workflow has to work across inconsistent account-management pages while preserving user approval and avoiding changes to the wrong account.

For security teams, the distinction between compromised, weak, and reused passwords is material. A compromised password is known or suspected to have been exposed, while reuse can create cross-service exposure even where an individual password has not been disclosed. Industry experience with browser-driven automation suggests that reliability depends heavily on each website's account flows, including MFA prompts, anti-automation controls, and password-reset rules.

Chrome's Canary channel is intended for experimental features and is not evidence of a release timetable. The source material does not provide a stable-release timetable for this Gemini-branded interface, identify which websites it would support, or explain how the model-assisted portion of the workflow would operate.