Google Chrome is now giving users more autonomy over the on-device AI models that strengthen its security features.

The company has quietly introduced a new option that allows users to delete the local AI models used by Chrome’s Enhanced Protection mode, a security feature that was upgraded with AI capabilities last year.

Enhanced Protection itself isn’t new—it has existed in Chrome for several years as part of Google Safe Browsing. Traditionally, the feature relied on server-side checks and known threat databases to warn users about dangerous websites, downloads, and browser extensions. However, in 2025, Google updated Enhanced Protection with AI-driven functionality designed to offer real-time threat detection.

While Google has not fully detailed how the AI-enhanced version differs from the older implementation, the company suggests that AI allows Chrome to detect suspicious behavior patterns on the fly. This could enable warnings for potentially harmful sites or files that Google hasn’t previously identified, rather than relying solely on known blacklists.

According to Google, the AI-powered system also performs more in-depth scans of downloads, potentially identifying malicious files before they can be executed on a user’s device.

What’s new is transparency—and control. Google has now confirmed that Enhanced Protection relies on an AI model stored locally on the user’s device. In the latest Chrome Canary build, users can remove these models entirely by navigating to:

Chrome > Settings > System > Turn off “On-device GenAI”

Disabling this option deletes the local AI model used for Enhanced Protection, effectively reverting Chrome to a more traditional security approach.

The change was first spotted by Leo and is currently limited to Chrome Canary, Google’s experimental release channel. However, the feature is expected to roll out to stable Chrome users in the near future.

This move comes amid growing scrutiny around on-device AI, privacy, and storage usage. By allowing users to delete local AI models, Google appears to be taking a step toward greater user agency—especially for those who are cautious about AI running locally or consuming system resources.


Buy ExpressVPN with PayPal or Credit Card

For now, Chrome users will soon be able to decide whether AI-powered protection is worth keeping—or whether they’d rather do without it.

Advertisement