Fake reviews: what changed in the rules in the EU, the UK, the US and Brazil

By Daniel Pisano, Partner & CGO at Vurdere
Fake reviews are an explicitly banned practice in the EU since May 2022, in the US since October 2024 and in the UK since April 2025. Fines reach $53,088 per violation in the US and 10% of global turnover in the UK. Brazil relies on its consumer code. The safe setup is the same everywhere: reviews from real buyers, published unfiltered.
Why are fake reviews a board-level topic again?
For years, buying a handful of five-star reviews to "kickstart" a new product was treated as a growth hack. Risky, maybe, but rarely punished. That changed quickly, and in several markets at once.
There is a second reason. Adobe reports that AI-referred visits to US retail sites grew 62% year over year in July 2026, and that those visitors convert 60% better than other traffic. Shoppers increasingly meet your reviews through a summary written by an AI assistant. I covered this in the guide to agentic commerce protocols, where the aggregate rating is a data field, not page decoration.
When the AI summary is the shopper's first impression, a fabricated review pollutes the source. It is now expensive, too.
What does the FTC rule ban in the US?
The Federal Trade Commission's Trade Regulation Rule on the Use of Consumer Reviews and Testimonials (16 C.F.R. Part 465) took effect on 21 October 2024. It groups the bans into six categories:
fake or false reviews and testimonials;
incentives conditioned on the review's sentiment;
undisclosed reviews by insiders such as employees or executives;
company-controlled review sites presented as independent;
review suppression, whether by threats or by filters that only publish the good ones;
fake indicators of social media influence.
Civil penalties can reach $53,088 per violation, according to law firm Benesch. In December 2025, the FTC sent warning letters to nearly a dozen companies asking them to confirm corrective steps in writing. Not a wave of lawsuits yet, but usually the signal that comes before one.
In my experience, the rule that catches retailers off guard is suppression. Plenty of teams filter out low ratings and call it moderation.
What has the EU required since 2022?
Directive (EU) 2019/2161, the Omnibus Directive, has applied across the EU since 28 May 2022. It did two separate things.
First, a disclosure duty: any trader that gives access to reviews must state whether, and how, it ensures they come from consumers who actually used or bought the product. It is a duty to be transparent, not a duty to verify every review.
Second, two practices went on the blacklist and are banned in all circumstances: claiming reviews come from real buyers without reasonable and proportionate steps to check, and submitting, commissioning or misrepresenting reviews to promote products. National rules add teeth. In Spain, Fieldfisher notes fines of up to €1,000,000 in serious cases.
More is coming. In June 2026 the European Commission announced the Digital Fairness Act as a legislative initiative for Q4 2026, with fake or incentivised reviews among its topics. The final scope is still being debated.
How far did the UK go?
Further than most. Under the Digital Markets, Competition and Consumers Act (DMCC), several fake-review practices became banned practices in April 2025, unfair and illegal by default. Where the CMA finds an infringement, it can impose fines of up to 10% of global turnover and require changes.
According to Taylor Wessing, the CMA swept more than 100 review publishers, sent advisory letters to 54 firms, and 90% of them made changes. In 2026 it opened five new consumer investigations. Google had already given the CMA formal undertakings on fake reviews in January 2025.
And Brazil, Vurdere's home market?
In my research for this piece, I found no Brazilian federal law written specifically about fake reviews. The framework comes from the Consumer Defence Code: Article 36 requires advertising to be identifiable as such, and Article 37 bans misleading advertising, including by omission. Marketplaces such as Shopee add their own bans on artificial interactions. A bill on e-commerce security (PL 3451/25) is moving through Congress, but the official summary does not mention reviews.
Market | Legal basis | Since | What weighs most |
European Union | Omnibus Directive (2019/2161) | May 2022 | Disclose how reviews are verified; buying reviews blacklisted |
United States | FTC rule (16 C.F.R. 465) | October 2024 | Up to $53,088 per violation; suppression banned |
United Kingdom | DMCC Act | April 2025 | Banned practice; up to 10% of global turnover |
Brazil | Consumer Defence Code (Arts. 36 and 37) | 1990 | Misleading advertising; marketplace rules |
What should your store change now?
All four rulebooks point the same way. If your operation meets the strictest, it meets the rest.
Ask every buyer for a review, not only the ones you expect to be happy. Choosing who gets asked is a quiet form of filtering.
Publish the negative ones. Moderation removes abuse, personal data and off-topic content under written criteria. It does not hide low ratings.
Explain on the page how reviews are collected. In the EU this is already mandatory: who writes them, how purchase is confirmed, what can be removed and why.
Decouple incentives from sentiment. If you offer a coupon, offer it for reviewing, never for reviewing positively, and say so.
Check agency contracts. If someone "generates reviews" for you, the liability is still yours.
Label insider reviews. Staff can review if it is clear who they are.
Where does Vurdere fit?
The safest way to get lots of reviews is to ask buyers on the channel they actually use. In Vurdere's research with more than 120,000 consumers, review requests sent by WhatsApp converted at up to 15%, against 5% for email. I wrote about it in WhatsApp for reviews.
In A/B tests with current clients, Vurdere collected up to 10x more reviews. Genuine volume fixes the problem that pushes stores to buy reviews in the first place: an empty product page. In the Vurdere app for Shopify, the aggregate rating and review count are written to the native metafields reviews.rating and reviews.rating_count.
My view
I think the fake-review debate is still stuck on legal risk, and legal risk is only half the bill.
The other half is technical. When an AI agent summarises reviews or compares two stores, it works with what it finds: volume, rating distribution, recency, wording. A 4.9 average built on forty near-identical reviews does not convince an attentive human. I see no reason it would convince a machine trained to spot patterns.
Europe has had rules since 2022, the US and UK have added real penalties, and Brazil will probably follow. A store that already runs to the strictest standard is covered either way, and better placed for shoppers who ask an AI before they ask a sales rep. Transparency in reviews is no longer a brand choice. It is an operating requirement.
Frequently asked questions
Does the EU require me to verify every review?
Not exactly. You must disclose whether and how you ensure reviews come from real buyers. Claiming they do without reasonable steps to check is what is banned.
Can I offer a discount in exchange for a review?
Yes, if the incentive does not depend on the review being positive and you disclose it. In the US, sentiment-conditioned incentives are banned under the FTC rule.
Can I delete negative reviews?
Removing abusive or off-topic content under published criteria is moderation. Hiding low ratings to lift the average is suppression, banned in the US and misleading in the EU and UK.
Is this legal advice?
No. It is an informational summary of public sources. For a specific case, talk to a lawyer.



Comments