Google Phone app adds fake call detection to flag AI scams

Scammers are no longer relying only on spoofed numbers. Google’s latest defense targets a more dangerous trick: calls that seem to come from someone already in a user’s address book, a tactic designed to exploit trust before a victim has time to think.

The new fake call detection feature in Phone by Google flags suspected spoofed calls when both the caller and the recipient are using Phone by Google. Google says the protection works automatically behind the scenes and is on by default. When a contact calls, the app uses a real-time device-to-device verification signal to check whether the call is likely legitimate, then marks suspicious calls so the user can hang up.

Google says the verification happens on-device, not in the cloud. No conversation audio or transcription is stored or sent to Google servers, and users can turn the feature off in settings or disable it for a specific call. The company is also cautioning that the system is not foolproof: it is not 100% accurate and will not catch every scam call.

The rollout is narrowly scoped. Google’s help center says Scam Detection works only on Google Pixel devices. In the United States, it is available on Pixel 6 and later, with Pixel 9a supported in the U.S. only. Pixel 9 and later devices support the feature in Australia, Canada, France, Germany, India, Ireland, Italy, Japan, Mexico, Spain and the United Kingdom.

The timing reflects how quickly fraud has scaled. Google cites INTERPOL’s March 2026 Global Financial Fraud Threat Assessment, which said impersonation fraud contributed to more than $400 billion in global losses. FTC data show impersonation scams caused $2.95 billion in losses in 2024. Google also said scammers used increasingly sophisticated tactics and generative AI tools to steal more than $1 trillion from mobile consumers globally in 2024.

The new call protection builds on Google’s recent verified financial calls feature, which warns users when a scammer is trying to impersonate a financial institution. It also extends a broader Android push that began in March 2025, when Google rolled out AI-powered scam detection for calls and texts, including message scanning in Google Messages for conversations with non-contacts and real-time warnings for suspicious conversational scams.

The practical advice remains unchanged even as the software gets smarter: do not trust a caller just because the number or name looks familiar. The FTC continues to warn that technology makes it easy for scammers to place millions of unwanted calls a day, and the safest response is still to block and report anything that does not pass a basic verification check.