Google Messages Introduces Gemini AI Scam Detection to Flag Fraudulent Texts Before Users Fall for Them

Google is expanding its scam protection features in Google Messages with a new AI-powered Enhanced Protection mode designed to identify suspicious text conversations before they can trick users into sharing personal information or money.

The update uses Google's Gemini artificial intelligence models to analyse conversations with unknown senders and recognise patterns that may indicate fraud. The feature is intended to address scams that develop gradually, with messages initially appearing harmless before the sender attempts to obtain sensitive information or persuade the recipient to make a payment.

Announced in October 2026, the enhanced safeguards are being rolled out gradually, giving Android users additional tools to identify increasingly sophisticated messaging scams.

How Gemini AI Identifies Suspicious Messages

Google Messages already includes scam detection that can flag suspicious messages using on-device machine-learning models. The new Enhanced Protection mode adds another layer by allowing cloud-based Gemini models to examine the broader context of conversations with unknown contacts.

This approach is designed to identify suspicious behaviour that may not be obvious from a single message. For example, a conversation might begin with an apparently routine enquiry before gradually shifting towards requests for money, account credentials or other sensitive information.

By analysing the context and intent of a conversation, Gemini aims to identify potential scams earlier than traditional filters that primarily recognise familiar patterns or known malicious links.

The feature supplements existing protections rather than guaranteeing that every fraudulent message will be detected. Users should continue to exercise caution when communicating with unfamiliar senders.

Gemini Nano Adds Protection Against Suspicious Links

The enhanced mode also introduces smarter checks for unfamiliar website links on supported devices.

Scammers frequently create new websites designed to imitate banks, delivery services or other legitimate organisations. Because these websites may not yet appear on established lists of malicious domains, conventional link filters can sometimes miss them.

Google's on-device Gemini Nano model uses real-time web signals to assess whether an unfamiliar link appears trustworthy before a user opens it. The feature is designed to provide another opportunity to identify potentially dangerous websites.

The update also introduces redesigned warnings for links received from numbers that are not saved in a user's contacts. These alerts provide additional context and encourage users to pause before opening a link.

Availability of the advanced link-checking capability depends on the device, with support initially focused on selected recent flagship smartphones.

Google Addresses Privacy Concerns

Using AI to examine message conversations raises questions about the privacy of personal communications. Google says Enhanced Protection uses its Private AI Compute technology to encrypt data and process it within a secure environment designed to prevent access by Google or other parties.

The cloud-based analysis supplements existing on-device detection. Users can choose whether to enable Enhanced Protection, while standard scam protection remains available separately.

The company says its existing AI-powered scam detection in Google Messages and Google's Phone app flags more than five million suspected scam calls and texts each month. This figure reflects the company's reported data.

Users who prefer not to use the enhanced cloud-based analysis can retain standard protection through the app's settings.

More Android Security Features Target Phone Scams

The messaging update forms part of a wider effort to strengthen Android's defences against fraud.

Google is also expanding scam alerts in notifications across additional Android devices. These warnings can help users identify suspicious messages in supported chat applications before opening them.

Other measures address attempts to misuse screen sharing and call forwarding. Android is extending automatic privacy protections to hide sensitive login and payment information during certain screen-sharing sessions. It is also introducing safeguards against unauthorised call forwarding, a technique scammers may exploit to intercept verification codes.

Together, these measures are intended to protect users across messaging, browsing and phone calls rather than relying on a single detection feature.

How to Enable Enhanced Protection in Google Messages

Users whose devices support the feature can check its availability by opening Google Messages, tapping their profile picture and navigating to Messages settings, followed by Protection and safety and Spam and scam protection.

Depending on availability, the menu may offer Enhanced Protection, Standard Protection and an option to disable protection.

The rollout is gradual, so the enhanced setting may not appear on every device immediately. Some link-checking capabilities are also restricted to supported models.

Google's latest update reflects the growing role of AI in mobile security. While the technology may help identify complex fraud attempts, users should still avoid sharing passwords, verification codes or financial information in response to unexpected messages.