Google Is Using Gemini to Know If You Actually Visited That Place Before Posting a Review

There's a quiet war being fought every time you search for a restaurant on Google Maps. On one side: businesses paying for fake five-star reviews, scammers extorting small shop owners with one-star bombs, and bots flooding listings with manufactured praise. On the other side: Google's Gemini AI, now deployed full-time as Maps' content cop — and it's getting very good at its job.

As of April 2026, Google has publicly confirmed a sweeping upgrade to how it handles trust and safety on Maps. The headline number from its 2025 Trust & Safety Report is staggering: over 292 million policy-violating reviews blocked or removed in a single year — alongside 79 million inaccurate or unverified edits blocked, 782,000 policy-violating accounts restricted, and over 13 million fake Business Profiles removed. That's not a cleanup. That's a full-scale operation. PPC Land

The Problem: Fake Reviews Are a Booming Industry

Google Maps recorded over 1 billion reviews and 80 million updates to business hours or contact information in 2025 alone. At that volume, manual moderation was never going to be enough. Bad actors figured this out long ago. heise online

One of the most insidious tactics that emerged in recent years is the review extortion scam — a bad actor posts a wave of fake one-star reviews, then contacts the business owner demanding payment, often hundreds of dollars, to remove them. Until recently, businesses had few good options: pay up, wait out a slow manual removal process, or watch their rating crater. Delete Facebook Account

Fake five-star reviews are equally common and arguably more damaging. They create a false sense of trust, drawing customers away from authentic businesses. Someone sees a 4.9-star locksmith, calls them for an emergency, and ends up being charged an outrageous fee for a job that should cost a fraction of the price. Leadtap

Enter Gemini: Three Layers of Defense

Google's response, announced on April 16, 2026, isn't a single fix. It's a three-layer system, with Gemini powering the core of it.

Layer 1 — Stop Scams Before They Go Live

Google has upgraded its systems to be faster at stopping new kinds of scams, like attempts to demand payment in exchange for removing fake one-star reviews. Its systems now better detect specific scam patterns, which lets it stop suspicious posts before they ever go live. If a sudden spike in spam reviews is detected, the fake content is quickly removed, new reviews on the profile are paused, the Business Profile owner is alerted, and a notification banner is displayed to let consumers know why contributions are temporarily paused. BetaNews

This is a meaningful shift — from reactive removal to pre-publication blocking.

Layer 2 — Gemini Reads the Reviewer, Not Just the Review

This is the part most people don't realise is happening. Gemini can analyze review behavior over time to spot unusual spikes. Even after a review is posted, the system keeps watching. Cybertegic

More importantly, Gemini doesn't just look at what a review says. It looks at who wrote it — and what that person's history suggests about whether they were ever actually there. AI is now being used to revisit older reviews to detect abuse patterns that weren't visible at the time of posting, which means fraudulent reviews can be removed even months later. Stan Ventures

Google screens every review with AI before it publishes, checking for spam patterns, policy violations, and suspicious behavior like bulk posting or identical language across submissions. A reviewer who has left 40 five-star reviews in the past two weeks across businesses in four different cities raises an obvious flag. Someone who has never visited a neighbourhood but leaves a detailed review of a hyperlocal café raises another. WiserReview

Layer 3 — Gemini Watches Profile Edits in Real Time

Gemini can now distinguish between a regular update — like a slight business name change — and a sudden, suspicious shift, such as changing a business category from "cafe" to "plumber." That kind of profile hijacking, often used to redirect customers to a scam operation, gets caught and blocked before it goes live on Android, iOS, and desktop globally. Search Engine Journal

A particularly notable case involved a network of scammers who impersonated locksmiths. They took control of unclaimed business profiles, presented themselves as legitimate service providers, and then charged customers exorbitant fees for emergency services. The incident served as both a warning to future bad actors and a training dataset for Gemini, making the model more effective at spotting similar scams in the future. Stan Ventures

The Reviewer History Signal Is the Real Game-Changer

The most technically interesting part of this update is how Gemini evaluates the reviewer, not just the review content. Unlike traditional rule-based systems, Gemini leverages machine learning to spot suspicious activity through nuanced patterns of behavior, and it only needs a few examples to start recognizing threats on a larger scale. Stan Ventures

What does that look like in practice? The system analyses patterns such as: Has this account reviewed many businesses in a short time? Do the reviews follow a template? Is the account newly created? Does the reviewer's location history suggest they were ever physically near the place they're reviewing? Does their overall contribution pattern match that of a genuine Local Guide, or someone running a review farm?

Google is also now making contributor reputation more visible — high-level contributors have gold-coloured profiles, making it easier to spot trusted voices, while achievement badges distinguish top reviewers, photo experts, and new contributors. The underlying signal from years of authentic contribution now carries real weight in how a review is evaluated. Android Authority

Business Owners Get a Seat at the Table

Beyond the automated systems, Google is rolling out proactive email alerts so verified and active business owners have a way to easily review important edits to their Business Profiles before they go live. For years, a bad actor could quietly alter a business's phone number or address and owners wouldn't notice until customers were already complaining. That changes now. Google

On April 17, 2026, Google also updated its Maps user-generated content policy, adding two new clauses under the Rating Manipulation section: no directing staff to hit review quotas, and no directing staff to solicit reviews that include specific content like an employee's name. Common agency tactics that had been grey areas for years are now explicitly off-limits. ReplyOnTheFly

What This Means If You Own a Business

The system is getting sharper, and it's catching more than just obvious fakes. Review deletion rates reportedly surged by 600 percent between January and July 2025 after Google integrated Gemini into the moderation stack. Some legitimate reviews got caught in the net too — short five-star reviews, clusters arriving close together, reviews following a template. ReplyOnTheFly

The practical takeaway is straightforward: ask all customers for honest feedback rather than selectively, avoid scripts or naming specific staff, respond to reviews to signal genuine engagement, and report suspicious activity through Google's reporting tools. Businesses operating with authenticity are better positioned as the system pushes out those gaming it.

The Bigger Picture

Gemini now works on both the presentation layer and the integrity layer of Maps data — surfacing business information to users through conversational AI on one side, and quietly filtering what information reaches those profiles on the other. A fake review doesn't just mislead someone on the Maps listing page anymore. It can pollute what Gemini tells users when they ask for a restaurant recommendation directly in the AI assistant. PPC Land

That's why Google is investing so heavily here. The integrity of Maps data is now the integrity of its AI answers. And with Gemini smart enough to know whether a reviewer has ever been near the place they're rating, the era of the paid review farm is looking increasingly short-lived.