How to Spot Fake Reviews in 2026: 10 Signs That Hold Up

Fake reviews are not rare and they are not all obvious. Most of them are glowing, many are written well, and the newest are written by AI. The good news is that the signs which hold up are about behaviour, and you can check them in a few minutes. This guide lists the ten that research and the platforms themselves rely on, what the law in the US, EU and UK now says, and how to report a fake review about your business. Every figure links to its source at the end, read on 21 October 2026.

How common are fake reviews?

The platforms publish what they catch, which gives a floor rather than a total. Google blocked or removed more than 292 million policy-violating reviews in 2025 and took down more than 13 million fake Business Profiles. Yelp received about 22 million reviews in 2025, recommended 70%, did not recommend 17% and removed 11%, and filtered nearly 500,000 suspected AI-written reviews. Trustpilot removed 4.5 million fake reviews in 2024, 7.4% of everything submitted, and 3.4 million of those were 5-star against 627,000 at 1 star. Fake reviews are mostly praise, so a page full of 5-star reviews is where to look hardest.

10 signs of a fake review that hold up

1. A burst of reviews. Many reviews in a few days, several on the same date, or a wave of praise right after a bad review. The FTC's consumer guide names bursts, and the UK regulator's guidance lists spikes as a detection signal.
2. Reviewer histories that do not fit. An account with one review ever, one that rates everything 5 stars, or one that reviews businesses on three continents in a week. In a study of reviews Yelp's filter removed, 75% of their authors had posted six or more reviews in a single day (Mukherjee et al., 2013), and behaviour signals beat wording signals.
3. Copied wording across reviewers. The same phrases, the same structure or the same odd spelling from different names. Rings of reviewers who all review the same businesses are a classic pattern; Yelp flags pages with suspicious activity such as many reviews from one IP address.
4. Only extremes. Reviews Yelp filtered as suspicious were more extreme than the rest (Luca and Zervas, 2016), and about 16% of Yelp restaurant reviews were filtered. Expect some fakes to give 4 stars to look believable.
5. Little concrete detail. Genuine hotel reviews described the room and the location, while invented ones talked about the occasion and the reviewer (Ott et al., 2011). Treat this as a weak sign: the same study found people judging by eye did little better than chance.
6. Details that do not match the business. A product the shop does not sell, a branch that does not exist, or reviews of a different item merged onto a listing, which the UK regulator calls catalogue abuse.
7. Gifts, refunds or free products. "They gave me a gift card" or "I was asked to update my review". Paying for positive reviews is banned in the US, UK and EU, and Yelp puts compensated-activity alerts on businesses it catches.
8. A rating that jumps, then sinks. Products that bought reviews got a short rating boost; once the buying stopped, ratings fell and the share of 1-star reviews rose (He, Hollenbeck and Proserpio, 2022). A recent rise that does not match older reviews deserves a closer look.
9. A platform warning. Yelp shows consumer alerts, Trustpilot shows consumer warnings and hides the TrustScore, and in the UK Google shows warning alerts on profiles caught boosting ratings. Check the top of the profile before the reviews.
10. No word on verification. Verified-purchase labels help. In the EU, a site that shows reviews must say whether and how it checks that reviewers really used the product.

Why "it sounds like ChatGPT" is not a sign

In a 2024 study, people could not tell reviews written by GPT-4 from real ones, even when paid for accuracy, and AI detection tools were fooled as well (Kovács, 2024). The FTC says spotting a fake by reading it can often be nearly impossible. Use signs 1, 2, 3 and 8, which are about behaviour, and treat writing style as a hint at most. Our free fake review detector checks a review against several of these signals; it gives an estimate, not a verdict.

What the law says now

This is a summary, not legal advice. If a decision depends on it, ask a lawyer in your country.

United States. The FTC rule on consumer reviews and testimonials (16 CFR Part 465) took effect on 21 October 2024. It bans fake or false reviews, including AI-written ones, buying reviews or paying for a particular sentiment, undisclosed reviews from officers and managers, review sites that pretend to be independent, suppressing reviews with threats or false accusations, implying the reviews you show are all of them when negative ones were hidden, and buying fake followers or views. The FTC can seek up to $53,088 per violation; that 2025 amount still applies because the 2026 adjustment was cancelled. It sent warning letters to ten companies in December 2025, settled with supplement maker TruHeight in 2026 for $750,000 over employee-written 5-star reviews and reviews traded for free products, and with Illinois sued Premium Home Service in May 2026 over thousands of fake listings and reviews (that case is pending).

European Union. Since May 2022 the Unfair Commercial Practices Directive, as amended by the Omnibus Directive, bans submitting or commissioning fake reviews and claiming reviews come from real customers without reasonable checks, and it requires sites to say whether and how they verify reviews. In a 2022 sweep of 223 sites, the Commission found 144 could not show they did enough to ensure reviews were genuine. In coordinated cross-border actions, maximum fines must be at least 4% of annual turnover.

United Kingdom. Since 6 April 2025 the Digital Markets, Competition and Consumers Act 2024 bans fake reviews, hidden incentives, publishing reviews in a misleading way (such as hiding negative ones or giving positive ones more prominence) and publishing reviews without reasonable steps against fakes. The CMA can fine up to 10% of turnover. In March 2026 it opened investigations into Autotrader, Feefo, Dignity, Just Eat and Pasta Evangelists.

If you show reviews on your own website

The same rules reach the reviews you display, not only the ones you collect. Three habits keep you on the right side of them. Show the platform's real overall rating and total number of reviews next to any selection you make, so a visitor can see the whole picture. Link to the full profile on Google, Trustpilot or wherever the reviews live. And never edit a review's words or rating. If your widget filters by star rating (ours offers "5 stars only" and "4 stars and up"), the first two habits matter most. More on showing reviews well is in the best review widgets for websites.

How to report a fake review about your business

• Google. In your Business Profile open Read reviews, choose the report icon on the review and pick a reason, or use Google's Reviews Management Tool, which also shows the status of each report. You get one appeal covering up to 10 reviews, and checks typically take several days. If someone demands money to stop posting fake 1-star reviews, do not pay; keep screenshots and use Google's merchant extortion form.
• Trustpilot. In Trustpilot Business go to Manage reviews, Service reviews, open the review and use the flag icon. For "not based on a genuine experience", Trustpilot asks the reviewer for proof; they have 3 days to answer before it goes to the content integrity team. The review stays up while it is checked unless it is harmful or illegal. Flag suspicious 5-star reviews too.
• Yelp. In Yelp for Business open Reviews, use the three dots on the review, choose Report Review and a reason. Yelp says it does not take sides in factual disputes, so report policy breaches, not disagreements. Menus change; check the current steps in Yelp's help centre.
• Facebook. Reviews are now called Recommendations. Open the recommendation on your Page, use the three dots and report it. Check Meta's help centre for the current menu.
• In the US, you can also report fake reviews to the FTC at ReportFraud.ftc.gov.

What not to do when you ask for reviews

Do not pay or reward people for positive reviews, ask staff or relatives to post them, or write any yourself. Do not hide or bury negative reviews while presenting the rest as the full picture. And check each platform's own rules, because some are stricter than the law: Google bans every incentive for a review and asking only happy customers, while the FTC allows incentives that do not depend on what the review says. Ask every customer, the same way, and let the reviews be what they are. For wording that stays within those rules, see our review request templates.

Sources

All read on 21 October 2026.

• FTC, Trade Regulation Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465
• FTC, Consumer Reviews and Testimonials Rule: Questions and Answers
• FTC, civil penalty amounts, 16 CFR 1.98
• FTC, warning letters to 10 companies under the review rule (Dec 2025)
• FTC, action against TruHeight (Apr 2026)
• FTC and Illinois, action against Premium Home Service (May 2026)
• FTC, How to evaluate online reviews
• FTC, How to report suspicious online reviews
• EU Directive 2019/2161 (Omnibus), amending the Unfair Commercial Practices Directive
• European Commission, sweep of 223 review websites (Jan 2022)
• UK Digital Markets, Competition and Consumers Act 2024, Schedule 20
• UK CMA, Fake reviews guidance (CMA208)
• UK CMA, 5 businesses under investigation over reviews (Mar 2026)
• Google, Maps user contributed content policy
• Google, protecting businesses on Maps (Apr 2026)
• Google, Report inappropriate reviews on your Business Profile
• Trustpilot, Trust Report 2025
• Trustpilot, How to flag service reviews from your business account
• Yelp, 2025 Trust and Safety Report
• Luca and Zervas, Fake It Till You Make It, Management Science (2016)
• Mukherjee et al., What Yelp Fake Review Filter Might Be Doing?, ICWSM (2013)
• Ott et al., Finding Deceptive Opinion Spam by Any Stretch of the Imagination, ACL (2011)
• He, Hollenbeck and Proserpio, The Market for Fake Reviews, Marketing Science (2022)
• Kovács, The Turing test of online reviews, Marketing Letters (2024)