Google Chrome is testing a new option that allows users to delete the on-device AI models used for scam and threat detection, giving consumers more granular control over browser security features.
Read More: Over 1.7 million users at risk from malicious chrome extensions
The change was first spotted in Chrome Canary by browser watcher Leo (@Leopeva64) and appears to be linked to Chrome’s “Enhanced Protection” mode — a feature updated last year to include real-time defense powered by local AI. These models analyze browsing patterns and flag phishing sites, malicious downloads, and suspicious extensions even before Google’s cloud systems identify them.
The addition of on-device intelligence represented a step forward for Chrome’s security stack, enabling faster and more private analysis that does not require constant server communication. Google has said the feature uses AI to detect threat patterns dynamically, though it has not fully disclosed how its methodology differs from earlier versions of Enhanced Protection.
The latest tweak introduces the ability to disable and remove the AI component altogether. Users can navigate to Chrome Settings > System and toggle off “On-device GenAI,” which triggers the deletion of the local model from the device.
While the capability is currently limited to experimental Canary builds, developers indicate it could ship broadly in future Chrome releases as part of a wider trend toward user-controlled AI settings. Reports also suggest that these local models may later support features beyond scam detection, potentially touching productivity and personalization tools across Chrome’s interface.
Read More: OpenAI eyes chrome amid google monopoly trial
As major browsers ramp up AI-assisted security, Google’s decision to make the feature optional aligns with increasing scrutiny over device-level AI, data privacy, and user consent. The update also comes amid rising competition from Apple, Microsoft, and Mozilla, each exploring AI-driven enhancements with varying privacy trade-offs.