Google is rolling out new scam protections for Chrome that use a server-side Gemini model to detect deceptive websites and warn users when they copy suspicious commands from ClickFix attacks, the company announced this week.
The changes are part of Chrome's Enhanced Protection mode in Safe Browsing, which Google says makes users ten times safer from phishing, social engineering, notification abuse and other scams compared with Standard Protection.
A new "suspicious site" warning will appear for pages that show deceptive patterns but have not reached the confidence level required for Chrome's existing full-page red warning. "Suspicious site warnings notify users when a page exhibits deceptive patterns commonly associated with known malicious sites," Google explained. "By tailoring the warning UI in this way, we are able to safely expand our coverage to display a higher volume of protective warnings."
Gemini's role has shifted from on-device to server-side detection for Enhanced Protection users. Google says the model examines signals such as the brand represented on a page, its intent, and how likely it is to be a scam. Safe Browsing then combines that analysis with domain reputation and other security signals.
The company said Chrome minimises data shared with the server by only analysing page text when a site exhibits high-confidence scam indicators. "And whenever page content is sent off the device, it is processed ephemerally – meaning it is never logged, retained or used to train models," Google noted.
Enhanced Protection can also detect when a user copies a suspicious command from a ClickFix scam, a technique that tricks users into running malicious PowerShell or Command Prompt commands themselves. If Chrome suspects a ClickFix script, it removes personally identifiable information from the clipboard payload and sends it to Safe Browsing for analysis. If the analysis returns a dangerous verdict, Chrome displays a warning before allowing the user to proceed.
Analysis
Why This Matters
- The move to server-side AI analysis represents a significant upgrade in Chrome's ability to detect sophisticated scams in real time, potentially protecting millions of users from credential theft and malware.
- ClickFix attacks are a growing vector: by intercepting copied commands before they are executed, Chrome addresses a technique that bypasses traditional browser defenses.
- The privacy safeguards, ephemeral processing, no logging or training data retention, attempt to balance improved security with user data concerns, though the shift to server-side analysis will inevitably raise trust questions.
Background
Google Chrome has long offered Safe Browsing protections, with Enhanced Protection mode introducing more aggressive scanning and sharing of page data. The addition of Gemini models for scam detection began on-device last year, targeting tech support scams. This update moves detection to Google's servers, leveraging a more powerful model capable of analysing page intent and brand representation. ClickFix scams, which trick users into pasting malicious commands, have become increasingly common, prompting this new clipboard-based detection mechanism.
Key Perspectives
Google: The company positions the changes as a natural evolution of Chrome's security, emphasising that server-side analysis enables stronger protections without compromising privacy. The tenfold safety improvement claim reinforces Enhanced Protection's value proposition.
Privacy advocates and users: Some may worry about page content being sent to Google's servers, even with the ephemeral processing guarantee. The distinction between Enhanced Protection (opt-in) and Standard Protection (default) means only users who explicitly enable the feature are affected, but clear communication about what data is analysed will be crucial for trust.
Security researchers: The addition of ClickFix detection addresses a specific and hard-to-defend attack surface. However, the 37% prevention rate for actions after initial credential theft (from the Blue Report 2026, cited in Google's announcement) underscores that initial access blocking remains only part of the challenge.
What to Watch
- User adoption of Enhanced Protection mode and any changes to default settings in future Chrome releases.
- The effectiveness of the suspicious site warning tier: whether it reduces false positives while catching more malicious sites.
- Any reports of bypass techniques or privacy concerns related to the server-side Gemini analysis, particularly from security researchers.
