Google announced the rollout Thursday (May 8) of a fresh suite of artificial intelligence (AI)-powered safety features across its platforms to protect users from increasingly sophisticated scams.
The company has integrated its on-device large language model (LLM), Gemini Nano, into Chrome’s Enhanced Protection mode on desktop. This model analyzes websites in real time to detect and warn users about threats such as tech support scams that use pop-ups and full-screen takeovers. On-device processing lets the browser catch cloaked sites that hide their true content from traditional web crawlers.
On Android, Chrome now includes AI-powered notification alerts. When a notification appears suspicious, users get a warning with the option to unsubscribe or view the blocked content. Users can also dismiss the alert if it seems incorrect.
Google Messages and Phone by Google have featured since March on-device Scam Detection for texts and calls. These tools scan for scam-like behavior in SMS, MMS, RCS messages and voice calls, alerting users to act in real time. Google says all processing stays local on the device, ensuring user privacy.
Google’s defenses have evolved with its latest update. The company says its AI now blocks 20 times more scammy websites than it did three years ago. That increase comes from better detection of coordinated scam networks and support for multiple languages. Google also pointed to seeing a rise in misleading pages mimicking official resources including government services. In 2024, it implemented new protections decreased scams impersonating official sites by more than 70%.
PYMNTS has tracked Alphabet’s broader AI strategy, which includes embedding Gemini into Search and productivity tools and expanding its cloud services. These updates show how Google uses AI not just for innovation, but as a defensive moat around its users and brand.
By embedding AI protections directly into its services, Google aims to preempt scams at the point of contact — before users fall victim.
Source: https://www.pymnts.com/
