Google is adding a stronger scam shield to its default texting app. Google Messages for Android is starting to roll out an opt-in Enhanced Protection mode that uses Gemini models, both in the cloud and on the phone, to catch suspicious conversations and links that slip past ordinary spam filters. Google announced the mode on October 14 as part of a wider set of Android anti-scam features, and 9to5Google spotted it going live in the Messages beta on October 23.
What Enhanced Protection adds
Until now, Messages offered one level of defence that blocks spam and flags known harmful links. The new mode sits above it and switches on three extra layers:
- Cloud-powered Scam Detection: when you're chatting with an unknown sender, Messages can send the conversation to Google's more capable cloud Gemini models to judge the sender's intent. Google says this catches new, slow-burn scams that on-device checks can miss. The analysis runs in Private AI Compute, where Google says data is encrypted and inaccessible to anyone else, Google included.
- Smarter link checks: on-device Gemini Nano combines real-time web signals to decide whether an unfamiliar link is safe before you tap it. This targets freshly registered scam sites that aren't on blocklists yet.
- Link checkpoints: redesigned warnings slow you down before you open links from numbers not saved in your contacts. Google says this was previously limited to some countries or to phones using Android Advanced Protection.
Which phones get it
Availability isn't the same for every part. The Gemini Nano link checks need newer hardware: Google lists the Pixel 10 and later, the Galaxy S26 series, the Galaxy Z Flip8 and Z Fold8 and "other select devices". The link checkpoints, by contrast, are listed for phones running Android 6 or later. Google said the mode would arrive "in the coming weeks", so expect a gradual rollout.
How to turn it on
- Open Google Messages and go to Messages settings.
- Tap Protection & safety.
- Choose Enhanced protection. The other choices are Standard protection and No protection, which Google doesn't recommend.
9to5Google found the option in beta version 20261020_00_RC00, so stable users may need to wait a little longer.
The rest of Google's anti-scam push
- Screen-share privacy: Android already hides sensitive data in its own apps during screen sharing; now Chrome and Android WebView will blank out login fields and payment pages too. Google says this is rolling out to Android 15 and newer.
- Call-forwarding protection: apps will no longer be able to quietly turn on call forwarding, a trick used to steal one-time codes. Android will also warn you if you're about to type a forwarding code. This arrives in early December on Pixel and select phones.
- Chat scam alerts on more brands: notification-level scam warnings from Pixel are coming to phones from makers including Xiaomi, OPPO, vivo, OnePlus, Honor and Motorola later this year.
- Verified bank calls: working with banks such as Banco do Brasil and Itaú Unibanco, Android hangs up on calls that fake a bank's number. Google says more than 3 million spoofed calls have been blocked since June.
What it means
Scams have moved from obvious "you won a prize" texts to long, friendly conversations that only turn malicious later. On-device AI alone struggles with that context, so Google is bringing in its bigger cloud models, but only when you opt in. That opt-in design matters: sending message content to the cloud is a real privacy trade-off, even with Private AI Compute's promises. Google's figures also come from studies it commissioned, which is worth keeping in mind.
Our take
If you or a family member often get texts from unknown numbers, Enhanced Protection is worth switching on once it reaches your phone. The link checkpoints alone should stop many accidental taps. Owners of older or budget phones won't get the full feature set, so the usual rule still applies: don't trust links from strangers, whatever your phone says.
Sources: 9to5Google, Google Blog
