Android scam protections use cloud and on‑device AI to spot fraud

Google is adding an AI-powered shield to Android to counter increasingly clever scam techniques. The company says scammers now use spoofed numbers, fake bank alerts, and even AI-based voice cloning, making fraud harder to spot. The new protections target both Messages and call-forwarding abuse, aiming to stop scammers before they can trick a user into sharing sensitive data.

In practice, Android will use cloud-based analysis built on Gemini-backed technology to examine conversations with unknown senders for red flags. This analysis runs in a privacy-conscious environment described as Private AI Compute, which keeps user data safe while the system looks for suspicious patterns. In addition, a lighter on-device model called Gemini Nano will run checks on unfamiliar links in real time, using current web signals to flag risky sites before you click.

The new mode is optional at launch. Users can enable it for extra protection within Google Messages as part of Android’s broader scam defenses. Google notes that its existing detection already flags more than 5 million suspected scam calls and texts each month, and the new layers should catch even more threats.

Beyond Messages and call etiquette, Google is widening anti-scam protections to other parts of Android. On-device scam detection in chat notifications will reach hundreds of millions more users later this year. Chrome and Android WebView will also gain better masking of sensitive information during live screen sharing. Google is, meanwhile, tightening how call-forwarding codes work to curb stealth abuses and expanding its verified bank calls system.

Several banks have already benefited from Google’s verified bank calls system. Since June, the system has blocked over 3 million fraudulent calls for Banco do Brasil and Itaú Unibanco, with Revolut and other banks supported as well. Google says the goal is to make it harder for scammers to impersonate trusted brands and numbers to steal data or money.

What this means for everyday users is straightforward: if you want stronger protection, keep your Android device and apps updated and flip on the new protections in Google Messages. Always verify the caller’s identity and never share credentials or one-time codes unless you started the contact yourself.

From a tech perspective, Google frames this as a layered defense. Cloud-based analysis helps catch complex patterns across conversations, while the on-device checks provide quick warnings before you open a dangerous link. It won’t stop every scam, but it raises the bar for attackers who rely on familiarity and urgency to succeed.