Android deepfakes voice cloning scam detection RCS

Google's New Android Feature Catches AI Voice Clone Scams Before You Get Fooled

3 June 2026

Google is rolling out a new Android security feature called "fake call detection" to combat AI-powered deepfake scam calls, available from June 2026 on Android 12 and later devices. The feature uses the RCS standard to send encrypted verification signals between devices when a call is made, alerting recipients with an on-screen warning if the signal is missing and the call cannot be authenticated — indicating a likely spoofed or AI-cloned impersonation attempt. The move comes amid growing impersonation fraud losses, with the FTC reporting $2.95 billion in US losses in 2024 alone and INTERPOL linking such fraud to over $440 billion in global losses.

Google is rolling out a feature for Android that tackles one of the nastier fraud trends going right now: scammers using AI-cloned voices to impersonate people you actually know.

Called "fake call detection," the feature lands on Android 12 and above this month, starting with Pixel hardware, switched on by default. No configuration required.

The mechanism is straightforward enough. When someone in your contacts rings you, their device sends a silent encrypted confirmation over RCS. Your phone receives that signal and knows the call is legitimate. If the signal is absent, your device pings your contact's actual handset directly to ask whether it's making a call right now. If the answer is no, you get an on-screen warning telling you to hang up.

The whole exchange happens in real time, which matters because voice-cloning scams work precisely by catching people off guard in the moment.

There's a catch worth noting: this only functions when both parties are running Phone by Google with RCS enabled, plus the Contacts and Google Messages apps. If the person calling you is on some other dialler, the verification loop can't complete. Google has suggested anyone who wants the protection should install Phone by Google from the Play Store and set it as default.

The problem this is responding to is real and getting worse. The FTC recorded $2.95 billion in losses from impersonation scams in 2024. INTERPOL's threat assessment from March this year flagged impersonation fraud as a major contributor to over $440 billion in global fraud losses. These aren't edge cases.

The classic attack pattern combines two things: spoofing a contact's phone number so the caller ID looks familiar, and layering AI voice synthesis on top so the voice sounds right too. Caller ID alone was already unreliable before AI made audio forgery accessible to anyone with a laptop and a decent internet connection.

Google has been quietly building out this kind of protection for a while. Late last year it extended in-call scam detection to financial apps including Cash App and JPMorgan Chase's mobile banking platform. This latest feature broadens the net further, moving upstream to catch impersonation before a conversation even properly begins.

Whether widespread adoption happens quickly depends on how many Android users actually run Phone by Google rather than a manufacturer's own dialler. Samsung ships its own phone app. So do several other OEMs. The protection is only as good as its coverage.