Google Messages for Android is rolling out an opt-in Enhanced Protection mode that uses Gemini models to spot scams and check links, according to 9to5Google, which found the mode in the latest Messages beta. The mode replaces the previous on-off protection toggle with three choices: Enhanced, Standard and No protection. Scam Detection uses cloud Gemini models to judge a scammer’s intent in conversations with unknown senders, while link checks run on the device’s Gemini Nano model with live web signals. The rollout follows Google’s announcement of the mode earlier in October.
Three choices instead of a toggle
According to 9to5Google, the new mode is appearing with the latest Google Messages beta, version 20261020_00_RC00. To enable it, users go to Messages settings, then Protection and safety, where they now find three options rather than a single toggle. Enhanced protection is the new top tier, Standard protection is the existing level, and No protection is marked as not recommended. Until now, Google Messages offered Standard Protection to block spam and suspected harmful links. Enhanced Protection adds what Google calls advanced detection to prevent spam and scams, aimed at the more complex attacks that simple blocklists miss.
Because the mode is opt-in, it will only protect users who switch it on in settings. The beta sighting also leaves the wider rollout timetable unconfirmed: 9to5Google reports seeing the mode in the latest beta, but Google has not said in the report when every Messages user will see it. The device limits apply to the link checks specifically, since 9to5Google lists the supported phones for that feature while describing Scam Detection without a device list.
How the Gemini checks work
Scam Detection uses cloud Gemini models to evaluate complex message context and understand a scammer’s intent more quickly when the user is in a conversation with an unknown sender. According to 9to5Google, this approach is better than the existing on-device detection at identifying novel and sophisticated attacks. It uses Android’s Private AI Compute feature to encrypt messages and process them in a secure space that Google describes as private to the user and not available to anyone else, including Google. Enhanced link protections work differently: they use the on-device Gemini Nano model plus real-time web signals to judge whether an unfamiliar link is trustworthy before the user taps it. That link feature is available on the Pixel 10 and later, the Samsung Galaxy S26 series, the Z Flip 8 and Z Fold 8, and other select devices. The update also redesigns warnings to give users more time and context to pause before tapping links sent by numbers that are not in their contacts.
Why It Matters
Enhanced Protection matters because messaging scams have moved beyond obvious spam into conversations that look plausible until the trap closes. A system that reads the whole context of a conversation, rather than matching a single suspicious link or phrase, stands a better chance against attacks that are tailored to the recipient. Putting the heavier analysis in a private compute space also answers the obvious objection: scam checks should not require handing message contents to Google in readable form. The device limits on link checks are the catch to watch, since the strongest protection will reach newer phones first. For users on older devices, the redesigned warnings and the cloud scam checks still raise the bar, but the full package remains tied to hardware.
Background
Google has been pushing Gemini models into everyday products throughout 2026, from Gemini Notebooks becoming connected apps that users can query in chat to changes in which models Google AI Plus subscribers can choose. Messages is a natural next step, since it is where many users first meet a scam. Google announced Enhanced Protection earlier in October, and the beta rollout now shows how the three-tier setting will look on phones that receive it.
Source: 9to5Google
0 Comments