Fake reviews aren't a fringe problem anymore — they're a background hum on almost every marketplace, and generative AI has made them cheap to produce at scale. The goal here isn't to declare any single review "fake." You usually can't prove that about one review. What you can do is read the pattern: a listing's reviews, taken together, either behave like a crowd of real buyers or like a manufactured chorus.
The seven checks below run roughly fastest to slowest. Most of them take seconds and need no special tool — just a habit of looking one layer past the average rating before you trust it.
The 7 red flags
Read the wording, not the star rating
The star count is the easiest thing to fake; the sentences are harder. Fake and AI-written reviews lean on the same stock openers — "I was skeptical, but…", "as someone who…", "this was a total game-changer" — and they repeat the full brand and model name where a real buyer would just say "it" or "the charger." Flawless grammar across dozens of casual reviews is its own tell; genuine shoppers make typos and trail off. A five-star review that lists no specifics — no size, no delivery hiccup, no thing they'd change — is describing an idea of the product, not the product.
Check the timeline for review bursts
Sort by most recent and watch the dates. Authentic reviews dribble in as people actually receive and use a thing. A cluster of dozens of glowing reviews landing inside a 24–48 hour window, after weeks of silence, is the signature of a promo push or a paid campaign. The same pattern in reverse — a sudden pile of one-star reviews with no purchase behind them — is a coordinated review bomb. Neither burst tells you the product is good or bad; it tells you the ratings were gamed.
Open the reviewer's profile
One click on the reviewer name is the single most revealing check. A real account has a plausible shopping life — a coffee grinder, a phone case, a pair of running shoes over months. A review-farm account posts five or more "Verified Purchase" reviews in a single day across unrelated categories: power tools, collagen, a suitcase, a pet toy. Brand-new accounts whose entire history is five-star ratings for one seller's catalog are working, not shopping.
Read the rating distribution, not just the average
Click the star breakdown. A healthy product shows a spread — mostly fours and fives, a scatter of threes, some honest ones. Be suspicious of the J-curve: a wall of five-star reviews and a spike of one-stars with almost nothing in between. That shape usually means purchased praise colliding with genuine complaints, and the 4.6 average is an artifact of both. The middle of the curve is where real, mixed experiences live; when it's missing, the rating was manufactured.
Weigh the verified purchases
Filter to "Verified Purchase" reviews and see how much survives. Legitimate products typically carry a strong majority of verified reviews; a listing propped up by unverified five-stars deserves caution. Verification isn't a guarantee — sellers ship cheap filler items to generate a "verified" tag, and a review marked verified can still be paid — but the ratio is a useful signal. Then do the thing most shoppers skip: read the one- and two-star reviews first. Real negative reviews name a specific, reproducible failure — the strap tore in a week, the app won't pair — and repeated complaints across several buyers are worth more than any amount of praise.
Vet the photos
Reviewer photos are supposed to be the trustworthy part, which is exactly why farms fake them. Watch for the same unboxing shot uploaded by several different "reviewers," stock-looking images that don't match a real living room, and pictures with no dust, fingerprints, or lighting variation. Some fake photos are now AI-generated outright. Drop a suspicious image into a reverse image search to see where else it lives, and run it through an AI-image check if it looks too clean. Our guides on <Link href="/guide/how-to-reverse-image-search" className="text-blue-700 underline">reverse image search</Link> and <Link href="/guide/is-this-image-ai-generated" className="text-blue-700 underline">spotting AI-generated images</Link> walk through both.
Know the rules — but still verify yourself
There is a legal backstop now. The U.S. Federal Trade Commission's rule on fake reviews — finalized in October 2024 and in force since 2025 — bans fake or AI-generated reviews, reviews from people who don't exist, buying positive reviews, and suppressing negative ones, with penalties reaching tens of thousands of dollars per violation. That raises the cost of faking, but it doesn't clean up a listing in real time, and enforcement lags the fakes. Tools like Fakespot that once auto-scored reviews have shut down (Mozilla ended Fakespot in 2025), so the checks above — done by you, on the page — are still the reliable move.
The 60-second version
If you only have a minute: sort reviews by newest and check for a burst, click one gushing reviewer's profile, open the star breakdown to look for the empty middle, and read the two worst reviews for a specific, repeated complaint. Those four moves catch most manufactured ratings without any software. Consumer reporters have converged on roughly the same short test for exactly that reason.
Why the pattern beats any single tell
Any one red flag has an innocent explanation. A review burst might be a legitimate launch. Flawless grammar might be a careful writer. A J-curve might be a genuinely divisive product. That's why you weigh several signals at once — a burst of grammatically perfect five-star reviews from grab-bag profiles, on a listing with no verified purchases and reused photos, isn't four coincidences. It's a picture. Verification is about assembling the picture, then deciding, rather than trusting the number on top.
Not sure what's real?
FAXTR searches 100+ fact-checking organizations in one query — free, no login required. Spotted a coordinated fake-review campaign? Report it.