Google Search: AI to detect and block scams

Google has published the Fighting Scams on Search report ( PDF ) that explains how it uses artificial intelligence to detect and block millions of scams in search results. Additional protections are included in Chrome and leverage the Gemini Nano model.

AI to Block Scams in Search

Google has been using machine learning to detect scam sites and fake listings in Search for many years. In 2024, it further enhanced these systems by leveraging artificial intelligence, specifically Large Language Models (LLM). This has allowed it to proactively block hundreds of millions of scams per day.

By analyzing the text, Google identified the most common scams, such as those involving airline and government customer service. The number of scams decreased by up to 80%, as AI systems discovered that the sites and phone numbers were fake.

The Mountain View company emphasizes, however, that users need to be more careful. Google recommends verifying the domain (often similar to the legitimate one), reading the information about the site (by clicking on the three vertical dots next to the URL) and checking the site design for any oddities.

Better protections are also coming to Chrome 137 for desktop . Safe Browsing’s Enhanced Protection mode will use the Gemini Nano model to examine page content and flag fake sites.

Another feature will be available in Chrome for Android . Using an on-device machine learning model, the browser will detect notifications from fake sites. The user will then be able to block future notifications.

During the trial over remedies that should be imposed on Google to restore competition in the search engine market, an executive and CEO Sundar Pichai said that no other company can run Chrome and offer the same level of functionality, security and privacy.