Google Messages gets AI scam hunter

Google has introduced its new AI technology, Gemini, to detect fraud in Google Messages. This feature, called Enhanced Protection, reads suspicious conversations in the cloud and checks links before users open them. This new option will not be enabled by default and will soon be available for selected smartphones, such as the Pixel 10 and the Galaxy S26 series.

According to Google, the Enhanced Protection feature will perform three key tasks. First, conversations from unknown contacts will be sent to the cloud for analysis. This enables the AI to identify scams that gradually build up, such as an innocent chat that turns into an investment proposal. Second, unknown links will be pre-checked using current data from the web. Finally, users will have more time to respond to links, with an extensive warning if needed.

The cloud analysis takes place via Private AI Compute, meaning that the messages are encrypted and Google itself has no access to this information. This is a significant difference from the current scam detection feature, which operates entirely on the device itself. However, it is unclear whether this feature will also be effective for Dutch users, as fraud via SMS in the Netherlands is decreasing and primarily occurs via email and WhatsApp.

The Fraud Helpdesk recorded a significant increase in reports of phishing via email in 2026. Reports of shady phone and SMS messages are declining, while WhatsApp is increasingly being used for scams. This means that Google's new AI feature, while a step forward in the fight against fraud, may only address part of the problem, given the shift of scams to other platforms such as email and social media.

Read the full article from AI Insider.