You’re reading a text that looks official — “Delivery attempt failed, click here to reschedule” — and your stomach does a small, sensible flip. That exact moment is what Google is trying to intercept. Over the past week, the company quietly bolted scam-spotting into tools you may already use dozens of times a day: Circle to Search on Android and Google Lens on Android and iOS. Circle to Search now lets you literally draw a circle around suspicious text and get an AI-generated “overview” that explains whether the message smells fishy and why, without forcing you to copy, paste or leave the app where the message arrived.

Why bother? Because scams aren’t a niche problem anymore — they’re everywhere, and they’ve gotten craftier. Fraudsters send fake bank notices, bogus delivery links, phony job offers and impersonations that mimic logos and corporate tone perfectly. Google says this kind of fraud is so widespread that it’s built multiple layers of AI to fight it — systems that already block billions of malicious messages every month and which power protections across Gmail and Messages. The company’s push to fold scam checks into everyday search and imaging tools is essentially an argument about friction: people won’t open a safety app when their heart’s racing, but they will long-press their phone.

Here’s how the new Circle to Search check actually works in practice. On compatible Android phones, you long-press the home button or navigation bar to activate Circle to Search, then trace a circle around the text on screen. Under the hood, Google pulls wording, links and contextual cues and runs an AI analysis that surfaces an AI Overview overlay: a short, plain-language breakdown of whether the content looks like a scam and which elements triggered that assessment. It doesn’t just say “scam” — it highlights red flags (urgent payment demands, sketchy URLs, requests for personal info) and suggests immediate next steps like blocking the sender or reaching out to the company through an official channel. That overlay sits on top of your messaging app, so you don’t have to swap apps or paste anything into a search bar.

Google Lens gets the same logic for screenshots and messages that exist outside of regular SMS. If the risky message lives inside a social app, a website, a DM, or an ad, you can screenshot it, open Lens in the Google app, tap the screenshot and run the same analysis. That expands the feature’s reach past the world of standard SMS/RCS and into the messy cross-platform place where a lot of modern scams live. In short: if you can capture it, Google can try to explain it.

Privacy and trust are naturally the two sticky parts of any AI that inspects your messages. Google’s scam protections have trended toward on-device processing where possible — for example, many of the real-time scam checks in Google Messages are designed to run locally to protect privacy — and the company frames Circle to Search’s Overview as surfacing web evidence and signals rather than indexing every private conversation on its servers. Still, the distinction between on-device checking, web lookups and server-side processing matters to people, and Google says it only surfaces an AI Overview when it has “high confidence” in its assessments, a way of reducing noisy, low-value warnings.

This isn’t Google’s first rodeo in the scam-detection ring. The new Circle/Lens capability slots into a broader safety strategy that includes automated phishing filters in Gmail, scam warnings inside Google Messages, and an ecosystem of policies and tech aimed at choking off fraudulent ads and impersonations. Think of the new feature as a familiar interface — search and image lookup — turned into a quick safety check. The bet is that protective tech works better when it’s as frictionless as the risky thing it’s trying to prevent.

There are limits. No AI model is perfect: false positives could flag harmless messages, and false negatives might miss a novel scam tactic. Scammers also adapt — new domains and social engineering tricks will always test detectors — so this tool should be an added safety net, not a replacement for common sense. Standard hygiene still applies: don’t click unexpected links, verify account problems through official websites or verified phone numbers, and report or block suspicious senders. Google’s documentation and safety posts walk through these same recommendations and the ways you can report bad actors.

For now, the feature is rolling out as part of Google’s ongoing updates to Android and its apps; availability will depend on your phone model and region, and Google’s posts and follow-ups have the most reliable rollout details. If you want to try it: on Pixel and many modern Android devices, long-press the home or navigation button, circle the suspicious text, and read the AI Overview that appears — or take a screenshot and open it in Lens on Android or iOS. It’s a small step that aims to shave off a big slice of the cognitive load scams rely on: the panic and the hurry that push people to click before they check.

At the end of the day, this feels less like Google inventing a new safety product and more like it turning familiar tools into quieter safety rails. That’s not a cure-all, but it is the kind of practical design nudge that might save someone from a bad afternoon — or worse, a drained bank account. If enough people stop clicking, scammers lose the leverage they’ve been exploiting for a long time.