Fake reviews are the single most common form of shopping manipulation online, and they are getting harder to detect because generative AI now writes them at scale. The good news is that fake reviews still leave a predictable trail: repetitive language, thin reviewer profiles, unnatural rating patterns, and content that answers the marketing brief rather than the product. This guide gives you 12 signals to look for, the free tools that grade a page for you, and the specific safety checks to run before you trust a rating.
Quick answer
A review is likely fake if the reviewer has almost no history, the wording sounds like ad copy, the rating pattern spikes in short bursts, and the same phrases repeat across supposedly unrelated reviews. No single signal is proof on its own, but three or four together are a strong warning. Independent aggregators such as Fakespot and ReviewMeta can grade an Amazon, Sephora, or Best Buy product page in seconds, and platforms including Trustpilot and Amazon now publish public transparency reports that show how many fake reviews were removed each year.
Why fake reviews matter more in 2026
Consumer trust is the currency of e-commerce, and fake reviews damage that trust across the whole market, not just for the manipulated brand. A 2024 BrightLocal consumer survey found that 75 percent of shoppers say they have suspected at least one review on a business page of being fake. The World Federation of Advertisers has estimated that the global fake review economy is worth over 150 billion dollars a year in influenced spending. Regulators have responded: on 21 October 2024, the United States Federal Trade Commission Rule on the Use of Consumer Reviews and Testimonials came into force, allowing civil penalties of up to 51,744 dollars per violation for buying, selling, or suppressing reviews.
Platforms are pushing back too. Trustpilot reported that its systems removed 3.3 million fake or misleading reviews in 2023, up from 2.7 million the year before. Amazon says it blocked more than 250 million suspected fake reviews from reaching its site in 2023. The volume is huge, which means a lot still slips through, and that is where your own review-reading skills have to fill the gap.
What counts as a fake review
Not every negative or overly enthusiastic review is fake. The term covers a specific range of manipulation: reviews written by the seller or by people paid by the seller, reviews written by the seller against a competitor, reviews swapped in private Facebook or Telegram groups, reviews generated by AI without any real product test, and old reviews for a completely different product that are still attached because the seller changed the product but not the listing. A genuine review from a real buyer who happened to love or hate the product is not fake even if the tone sounds extreme.
12 signals that a review is likely fake
Use these as a checklist. Any one signal alone is weak. Three or four together are strong evidence.
- The reviewer has almost no history. A profile with one review, created the same week as the review, and no photo is a classic paid-review pattern.
- Bursts of five-star ratings clustered on a few days. Real reviews arrive slowly and steadily. A product that jumps from 20 to 200 reviews in three days almost always bought them.
- Wording that sounds like ad copy. Phrases such as “life-changing”, “a must-have for every home”, or “far exceeded my expectations” that repeat across several reviews signal a template.
- The reviewer names product features that are not in the listing. Suggests the writer had a marketing brief, not the product.
- Every review is written in the same style. Same sentence length, same opening line, same emoji, same sign-off across many reviewers.
- Very short reviews with a five-star rating. “Great!” or “Love it” alone, at high volume, is easy to farm and easy to fake.
- The reviewer has reviewed dozens of unrelated products in a short window. Fifteen five-star reviews for kitchen gadgets, dog food, USB cables, and mascara in the same week is a review-farm account.
- The negative reviews are all suspiciously similar. Coordinated attack reviews against a competitor also happen. Same complaint, same tone, same posting week is the tell.
- The rating distribution is bimodal. A product with almost no three- or four-star ratings and a wall of five stars plus a few one stars often has paid stars and honest disappointment.
- The reviewer got the product free. Incentivised reviews are legal on Amazon Vine, but they must be disclosed. Undisclosed “I received this product for free in exchange for an honest review” is a red flag and, in the United States, a violation of the FTC rule.
- The review talks about a completely different product. Sellers sometimes swap a product on an old listing to inherit its reviews. If most of the top reviews are about a garden hose and the listing now sells a phone stand, walk away.
- The wording feels AI-generated. Overly balanced structure, no personal detail, no specific use case, and no photo of the actual item are all signs of a language-model-written review.
Which review platforms are the easiest to game
All of them can be manipulated, but the effort required varies widely. Independent audits and platform transparency reports suggest a rough hierarchy: Amazon marketplace listings and small-shop Trustpilot profiles are the most heavily targeted because a rating change directly moves sales. Google Business Profile reviews on local services sit in the middle because Google now cross-checks reviewer location and account age. TripAdvisor and Booking.com hotel reviews are harder to fake because only verified guests can leave a rating on Booking.com and TripAdvisor uses stricter fraud detection. Professional B2B sites such as G2 and Capterra require LinkedIn verification, which cuts most casual review farming.
Free tools that grade a page for you
You do not have to do all the work by hand. Three tools cover the most common shopping scenarios in 2026:
- Fakespot (fakespot.com, now owned by Mozilla). Paste an Amazon, Best Buy, Sephora, or Walmart product URL and Fakespot returns a letter grade from A to F based on reviewer patterns. It is the most established tool in this space.
- ReviewMeta (reviewmeta.com). Focuses on Amazon and recalculates the star rating after stripping out suspicious reviews. Useful when Fakespot gives a low grade because ReviewMeta shows you what the honest rating looks like.
- The Transparency Company and platform-native transparency reports. Trustpilot, Amazon, TripAdvisor, and Google now publish annual fake-review numbers. Reading these tells you how aggressive the platform is at enforcement.
How to read a Trustpilot or Google Reviews page like a pro
Trustpilot and Google are the two review layers most people rely on for websites and local businesses. Both have specific tells worth learning:
On Trustpilot, sort the reviews by newest first, then filter to one-star reviews. Look for a wall of unrelated complaints (“my parcel never arrived”, “charged twice”, “cannot contact support”) repeating across many months. Then check the company response: legitimate companies reply with a case reference and an offer to resolve. Silence or copy-paste apologies at scale is a bad sign. Also check whether the business has “claimed” its profile and how many invited reviews versus organic reviews it has. A profile that is 95 percent invited reviews is closer to a marketing tool than an independent rating.
On Google, click on the reviewer’s name to see their full history. Google shows every review the account has left. A five-star Google review from an account that has posted three five-star reviews for three unrelated small businesses in the same city on the same day is almost certainly paid.
What to do if you spot a fake review
Report it. On Amazon, use the “Report abuse” link under each review. On Trustpilot, click the flag icon and select “suspicious activity”. On Google, use the three-dot menu next to the review to “Report review”. In the United States, systematic fake reviewing can also be reported to the FTC at reportfraud.ftc.gov. Platform-level reporting is far more effective than posting an angry comment because it triggers an automated re-check of the reviewer’s whole history.
Frequently asked questions about fake reviews
Is it illegal to write a fake review?
In the United States, yes. Since 21 October 2024, the FTC Rule on the Use of Consumer Reviews and Testimonials makes it illegal to write, buy, sell, or suppress reviews with civil penalties up to 51,744 dollars per violation. The European Union covers the same ground under the Omnibus Directive, which requires platforms to disclose how they verify reviews.
Can AI reliably detect AI-written reviews?
Not with high confidence in 2026. Most AI-detection tools produce false positives when tested on real human reviews. Pattern analysis of reviewer behaviour (account age, review velocity, product diversity) is still more reliable than trying to detect the text itself.
Do incentivised reviews count as fake?
Not automatically. Amazon Vine and similar programmes are legal in the United States as long as the reviewer discloses that the product was received free. A review that fails to disclose this is a violation.
Why do fake reviews sometimes look negative?
Because sellers also pay for negative reviews against competitors. This is called review bombing. The pattern is similar to positive fakes: burst timing, template wording, and reviewer accounts with no other activity.
Should I still trust product ratings at all?
Yes, but weight them differently. Prioritise reviews with a verified purchase badge, look at the shape of the rating distribution rather than just the average, read the three-star reviews (they tend to be the most honest), and cross-check the product on a second platform before you buy.
Sources and further reading
- Federal Trade Commission, Rule on the Use of Consumer Reviews and Testimonials, effective 21 October 2024, ftc.gov.
- Trustpilot, Transparency Report 2024, covering fake-review removals in 2023.
- Amazon, Brand Protection Report 2023, section on inauthentic review prevention.
- BrightLocal, Local Consumer Review Survey 2024, section on trust in online reviews.
- Fakespot (Mozilla), analyser tool, fakespot.com.
- ReviewMeta, adjusted-rating calculator for Amazon, reviewmeta.com.
- webwiki.com, How to Tell If a Website Is Trustworthy: The Complete Checklist.
- webwiki.com, How to Read Website Reviews Without Getting Fooled.
- webwiki.com, The Most Common Online Scams in 2026 and How to Avoid Them.