Google enhances Chrome, Search and Android with AI to crush online scams

In its latest “Fighting Scams in Search” report, Google highlights AI-driven, real-time protection measures embedded directly into devices and platforms to identify and neutralize fraudulent content and deceptive tactics before reaching users

Google has unveiled a sweeping set of updates powered by artificial intelligence to counter a sharp rise in online scams, introducing new tools across Google Search, Chrome, Android, and its communication apps. These measures, detailed in its latest “Fighting Scams in Search” report, reflect the tech giant’s intensified focus on safeguarding users through AI-driven, real-time protection embedded directly into devices and platforms. The updates aim to identify and neutralize fraudulent content and deceptive tactics before they can reach or harm users.

Google search scam detection improved

One of the core areas of improvement is Google Search, which has become a frequent target for scams that attempt to mislead users with fake support numbers, phishing links, and fraudulent service pages. Google reports that enhancements to its AI classification models have led to a twentyfold increase in the detection and removal of scam content. These models now better recognize patterns typical of coordinated scams and can identify subtle shifts in the language or structure of fraudulent websites. An example highlighted in the report involves fake airline support listings, a common trap for users looking for flight assistance. According to Google, these scams have dropped by more than 80% in Search results following the latest AI updates.

Gemini Nano enhances browser security

To further strengthen its browser security, Google is integrating Gemini Nano, its lightweight on-device AI model, into Chrome’s Enhanced Protection mode. Unlike traditional cloud-based threat detection, Gemini Nano analyzes web content in real-time directly on the user’s device. This approach offers more immediate protection against fast-evolving scams, such as tech support frauds that deliver alarming messages urging users to take hasty actions. Gemini Nano is initially focusing on detecting these scams within Chrome but is expected to expand to Android devices in the near future. By avoiding reliance on cloud processing, this model ensures quicker threat response while preserving user privacy.

AI screens malicious notifications

Another area where scams have grown increasingly sophisticated is through browser notifications. Cybercriminals have exploited notification permissions to deliver spam, fake alerts, and misleading content straight to user devices. In response, Chrome on Android will now feature AI-based notification screening. The system analyzes the behavior and content of websites requesting notification access. If the site appears untrustworthy, users receive a warning before proceeding, reducing the chances of falling for deceptive messages. An override feature is available for those who believe the AI flagged a site in error, maintaining user control over the experience.

On-device AI protects communications

Google is also reinforcing defences within its communication tools. AI-powered scam detection has been embedded into the Google Phone and Messages apps on Android. These systems work offline, relying on on-device AI to assess texts and calls for signs of fraud, such as urgent requests for personal information or impersonation attempts. By alerting users early in the interaction, the apps provide a critical layer of protection against increasingly subtle and manipulative scams.

Overall, these updates represent a significant shift in Google’s cybersecurity approach, focusing on localized, privacy-first protections that are responsive and adaptive. While no system can block every threat, the use of on-device AI offers a faster, more scalable method for identifying scams as they arise. Google plans to continue expanding these tools to more platforms and threat types, underscoring its commitment to staying ahead in the evolving landscape of digital fraud.