The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Google is adding two AI-assisted defenses to Chrome: on-device scam-page detection and automatic cleanup for suspicious website notifications. They extend Chrome’s existing Safe Browsing protections rather than replacing them with an all-purpose AI antivirus.
The scam detector is associated with Chrome 137 and Enhanced Protection on desktop. The notification feature helps users deal with fake virus warnings, bogus delivery alerts, fraudulent giveaways and other unwanted messages. Availability can vary by Chrome version, device, region, language, hardware and administrator policy.
As an Amazon Associate I earn from qualifying purchases.
The two AI features, at a glance
| Feature | What it does | Where it fits |
|---|---|---|
| Gemini Nano scam detection | Analyzes suspicious pages for signals associated with scams, especially fake technical-support pages. | Chrome Safe Browsing’s Enhanced Protection mode |
| AI notification protection | Identifies potentially spammy or scammy site notifications and offers to unsubscribe or show them. | Chrome’s website-notification controls |
Google has described several AI safety improvements for Chrome, so “two tricks” is best understood as a shorthand for these two related protections—not the name of a single product bundle. Other announcements have also mentioned help with compromised passwords, fake-virus pages and fake giveaways. See Google’s Chrome AI safety announcement.
Free tools Windows power users keep installed
One-click scans. No signup required.
1. Gemini Nano looks for scam pages
Traditional Safe Browsing is highly effective at blocking known dangerous URLs, but scammers can quickly create new domains or change the wording and design of a page. Google says Gemini Nano can help identify patterns that may not yet have produced a conventional bad reputation.
#1 Best Overall
A typical detection flow works like this:
- Chrome notices triggers associated with a possible scam, such as behavior linked to a keyboard-lock or fake-support page.
- The on-device Gemini Nano model examines the suspicious page.
- It extracts security-related signals, including the page’s likely intent.
- Those summarized signals are sent to Safe Browsing for Enhanced Protection users under Google’s described data-handling model.
- Safe Browsing combines the AI signals with other intelligence and metadata. If it concludes that the page is dangerous, Chrome displays a warning interstitial.
Gemini Nano does not independently declare every suspicious page malicious. It supplies additional evidence to Safe Browsing, which makes the final determination. Google says the model is triggered selectively and uses asynchronous processing, token limits, throttling and quotas to limit resource consumption. Google’s technical explanation is available in its description of Gemini Nano scam detection.
What it may catch
- Fake “your computer is infected” pages
- Full-screen browser takeovers
- Pages that try to lock keyboard or mouse input
- Fake Microsoft, Apple or Google support pages asking you to call a number
- Urgent account-security warnings and fake giveaways
- Pages demanding remote-access software, payment, gift cards or cryptocurrency
2. Chrome can clean up scammy notifications
Website notifications are legitimate when they come from a service you trust, but scammers often use notification permission to imitate antivirus alerts, operating-system warnings, delivery updates and prize messages.
Chrome’s AI-assisted notification protection can identify notifications that look suspicious, excessive or unwanted. Depending on the situation, it may offer two choices:
- Unsubscribe: stop notifications from that website.
- Show notification: display the particular notification despite Chrome’s warning.
This is a permission and nuisance-control feature, not proof that every hidden notification is malware. Google has reported that Android Chrome users received substantially fewer scammy and spammy website notifications with its AI-powered protections; that is Google’s aggregate claim, not an independent measurement of every user’s experience. Read Google’s broader Chrome AI and safety overview.
How to turn on Chrome’s strongest protection
The Gemini Nano scam-detection layer is associated with Enhanced Protection, Chrome’s highest Safe Browsing setting. Labels can vary slightly by release, but the usual desktop path is:
- Open Chrome and select the three-dot menu.
- Choose Settings.
- Open Privacy and security.
- Select Security or Safe Browsing.
- Choose Enhanced protection.
Enhanced Protection provides more proactive defenses against phishing, malware and scams than Standard Protection, but it also sends more browsing-related information to Google for analysis. Google’s Safe Browsing explanation describes the trade-off.
Rank #3
Standard versus Enhanced Protection
| Setting | Practical meaning |
|---|---|
| Standard protection | Baseline warnings against known dangerous sites and downloads, with less information shared than Enhanced Protection. |
| Enhanced protection | Stronger, more proactive detection, including AI-assisted signals for some novel scams, with greater browsing-related data sharing. |
| No protection | Chrome’s Safe Browsing warnings are disabled and is not recommended. |
Is the AI private?
Google says Gemini Nano runs locally on the device for the scam-page analysis. That reduces the need to send the page itself to a cloud AI service. However, “on-device” does not mean that nothing leaves your computer.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →For users who opt into Enhanced Protection, Google says summarized security signals can be sent to Safe Browsing. Enhanced Protection also involves sending more information about visited sites to Google than Standard Protection. In other words, local model execution limits some cloud-processing exposure, but it is not the same as private browsing or a no-data-sharing mode.
Does every Chrome user have it?
No universal rollout matrix has been established for every Chrome channel, operating system, country, language, hardware configuration and managed account. Google announced the desktop scam-detection feature alongside Chrome 137 in May 2025 and said it planned to expand detection to other scams, such as package-tracking and unpaid-toll scams, as well as to Android.
Do not confuse this background security model with Gemini in Chrome, the user-facing assistant. They are separate features and can have different eligibility, privacy and subscription requirements. Google’s Android Chrome announcement and Gemini in Chrome support page document that distinction.
Work and school browsers may also be controlled by an administrator. Chrome Enterprise administrators can manage Safe Browsing-related policies; users can check for restrictions at chrome://policy.
If the setting or feature is missing
- Update Chrome through Settings → About Chrome.
- Restart the browser and check Settings → Privacy and security → Security again.
- Open
chrome://policyto see whether an employer or school controls the setting. - Confirm that you are using the intended Chrome profile.
- Allow for staged rollout, hardware, language, country and platform restrictions.
Do not install a random “AI scam blocker” extension simply because the built-in option is unavailable. Some Chrome builds may expose local-AI controls under Settings → System or another AI settings section, but the exact label is release-dependent. Disabling local AI may remove or weaken features that depend on it.
What Chrome still cannot do
- False positives: a legitimate page or notification can be flagged.
- False negatives: new, carefully disguised or rapidly changing scams can evade detection.
- Compromised legitimate sites: a familiar domain can still be hacked or abused.
- Threats outside the browser: Chrome cannot prevent every malicious attachment, phone scam, extension problem or social-engineering attack.
- Human manipulation: a warning cannot stop someone from voluntarily giving a scammer remote access or payment.
Scammers also adapt their timing, content and techniques. Google has identified continuing challenges including prompt injection and attempts to bypass detection.
What to do when a browser warning looks alarming
- Do not call a phone number shown in a pop-up claiming to be Microsoft, Apple, Google or antivirus support.
- Do not install software because a page demands it.
- Never grant remote access to someone who contacted you through a browser warning.
- Close the tab or window rather than clicking controls inside a suspicious full-screen page.
- Check the real domain before signing in or entering payment details.
- Restore a legitimate notification deliberately through Chrome’s notification settings if it was hidden by mistake.
- Use strong, unique passwords, multifactor authentication, automatic updates and your operating system’s built-in security.
The bottom line
Chrome’s two most useful AI safety additions are Gemini Nano-assisted scam-page analysis and AI-assisted notification cleanup. They can add an extra layer against fake support pages, alarming browser pop-ups and abusive notifications, particularly when Enhanced Protection is enabled.
They are not a guarantee, and Enhanced Protection has a privacy trade-off because more browsing-related information may be sent to Google. Treat the features as defensive assistance—not as a replacement for updates, multifactor authentication, password hygiene and skepticism toward urgent demands.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
