Three million blocked fake calls since June. That number, buried near the end of Google’s latest Android security announcement, tells you more about the scale of the mobile fraud problem than any abstract statistic could. And it helps explain why Google is now moving Gemini from a chat assistant into the role of fraud cop.
Google announced a set of new and expanded scam protection tools for Android that touch messages, incoming calls, and screen sharing. The headline feature is a new opt-in setting in Google Messages that runs Gemini models against incoming conversations to evaluate whether a sender might be running a scam. It’s not just flagging known spam numbers. The system looks at the intent behind the messages themselves, including the conversational pattern of a scammer pretending to have the wrong number and then trying to establish contact anyway.
Privacy is the obvious concern when you’re sending message content to an AI for analysis. Google says Android processes messages inside Private AI Compute, an encrypted environment that neither Google nor outside parties can access. Gemini Nano also handles on-device link analysis in Messages, checking URLs before you tap them. That combination of on-device processing and isolated cloud analysis is a reasonable answer to the privacy objection, though it will still require user trust. The new scam warning setting starts rolling out in the coming weeks.
The device support list matters here. On-device Gemini Nano link analysis is coming to Pixel 10+, Samsung Galaxy S26 series, Galaxy Z Flip 8, and Galaxy Z Fold 8. Scam Detection in message notifications, which surfaces warnings without you even opening the message, is expanding beyond Pixel to devices from Honor, Motorola, OPPO, OnePlus, realme, vivo, and Xiaomi later this year. That’s a significant portion of the Android market, which makes this less of a Pixel perk and more of a platform-wide push.
Two other features round out the announcement. Automatic screen-sharing privacy rolls out later this month for all devices on Android 15 and above. It hides passwords, banking details, and card security codes by default when Chrome or Android WebView is visible during a screen share. This is a direct response to social engineering attacks where victims are manipulated into sharing their screens while handling sensitive information. And starting in early December, call forwarding protection arrives on Pixel and select other devices. It blocks apps from enabling call forwarding silently in the background and alerts users if a scammer tries to get them to manually enter an MMI code to reroute their calls.
- Gemini-powered scam detection in Messages, opt-in, rolling out soon
- On-device link analysis via Gemini Nano on Pixel 10+, Galaxy S26, and Galaxy Z series
- Scam Detection in notifications expanding to Honor, Motorola, OPPO, OnePlus, realme, vivo, Xiaomi
- Automatic screen privacy during screen sharing on Android 15+ devices, arriving later this month
- Call forwarding protection on Pixel and select devices, rolling out in early December
- Bank call verification in Brazil, with more partner banks planned
The Brazil bank call verification program is worth watching. Android already performs live verification checks on incoming calls claiming to be from partner banks, and hangs up automatically if the caller fails. Three million blocked calls in roughly four months is a meaningful number. Google hasn’t committed to bringing that to other markets yet, but the infrastructure is clearly there.
Apple has its own scam filtering in iMessage and has been tightening privacy around screen sharing, so this isn’t Google moving into uncontested space. But the Gemini integration, especially the conversational intent analysis rather than just blocklist matching, is a step beyond what most platforms currently offer. Whether users actually opt in, and how well the model handles edge cases without over-flagging legitimate messages, will determine whether this becomes genuinely useful or just another buried setting most people never touch.
