Fake reviews and AI testimonials: the FTC rule a small business has to follow
Since October 2024 a US federal rule bans fake and AI-written reviews, rewards for five stars, undisclosed staff reviews and hiding bad ones, at up to $53,088 per violation. What each part means for a small business, what Google and Trustpilot remove on their own, where the platforms are stricter than the FTC, and how to ask for honest reviews instead.
A review you bought, a testimonial a chatbot wrote for a customer who doesn't exist, five stars from your brother-in-law with no mention that he is your brother-in-law: for years these were bad manners that platforms sometimes cleaned up. In the United States they have broken a federal rule since 21 October 2024, with fines that can run to tens of thousands of dollars for each one.
The rule is the FTC's Trade Regulation Rule on the Use of Consumer Reviews and Testimonials (16 CFR Part 465). It applies to any business that markets to US consumers, whatever its size: a salon, a plumber, a Shopify store or a five-person software company. In December 2025 the FTC sent its first warning letters under it, to ten companies it did not name. This piece covers what the rule bans, what it costs to break it, what Google and Trustpilot now catch on their own, and how to ask for reviews the honest way.
What the rule bans
The Commission voted the final rule through 5-0 in August 2024. It has six prohibitions, and the FTC's own announcement names AI-generated fake reviews as an example of the first one.
Some things are left alone, and the FTC's Q&A is clear about them. A business that only hosts reviews is generally not liable for fakes it didn't write or buy. An email asking every customer for a review is a "generalized solicitation", so you are not on the hook if an employee who happens to be on the list writes one. Sorting reviews by helpfulness is fine. You may reply to a bad review in public and contact the customer to put things right.
AI testimonials: where the line is
The test is not whether a tool touched the words. It is whether the review is truthful about who wrote it and what they experienced.
- Banned: a review or testimonial that an AI made up and that you publish as if a customer said it. Also a "customer" quote on your landing page with a stock photo and a name you invented.
- Covered, not banned: an AI avatar or synthetic voice reading a testimonial. The FTC says avatars are not banned in themselves, but the testimonial is judged like any other: if it is fake or doesn't reflect a real customer's experience, it breaks the rule. A celebrity look-alike used without permission is a problem of its own.
- Fine, with care: a real customer's words, shortened or tidied, that the customer has seen and agreed to. Keep their written OK, and don't change what they said about results.
One twist from December 2025: the FTC set aside its 2024 order against Rytr, a writing tool with a "testimonial and review" generator, saying the complaint did not show a violation by the tool maker. That changes nothing for the business that posts the review. The ban on publishing fake reviews is untouched, and the warning letters went out the same month.
What it can cost
The rule matters because it lets the FTC ask a court for civil penalties for knowing violations, which it could not do under the general FTC Act alone. Since 17 January 2025 the maximum is $53,088 per violation (16 CFR 1.98, adjusted each year for inflation). Each fake review can count as a violation, so the ceiling grows fast.
Before the rule existed, the FTC already took $4.2 million from the clothing retailer Fashion Nova in January 2022 for holding back reviews below four stars that its review tool had collected. State attorneys general can bring their own cases. And anyone selling to the UK should know that since 6 April 2025 the Digital Markets, Competition and Consumers Act bans fake reviews, hidden negative reviews and undisclosed incentivised reviews outright, with fines the Competition and Markets Authority can set itself of up to £300,000 or 10% of worldwide turnover, whichever is higher.
The cheapest part of a fake review is buying it.
What the platforms catch before the FTC does
For most small businesses the first and likeliest penalty is not a lawsuit. It is the platform taking the reviews down, and sometimes telling your customers why.
Google's systems look for patterns on the account and at the place, such as a sudden burst of five-star reviews. Yelp puts a Consumer Alert on a business page when it finds paid or incentivised review activity. A wave of removed reviews often takes honest ones from the same week with it, and a public pause banner does more harm to a local business than a three-star average.
The platforms are stricter than the FTC
This is where owners trip. Something the federal rule allows can still get your reviews removed:
The safe rule of thumb: never reward anyone for a review on Google or Yelp, ask everyone or no one, and keep staff off your own listings.
How to ask for honest reviews
Honest reviews come from asking at the right moment, every time, in a way that takes the customer ten seconds.
1. Get the direct link. In your Google Business Profile, use "Ask for reviews" to copy the short link that opens the review box, and make a QR code from it for the counter or the receipt.
2. Ask everyone the same way. One message after every job or order, sent the same number of days later, whatever you think of the customer. Sending it only after a compliment is gating.
3. Keep it short and neutral. For example: "Thanks for coming in on Tuesday. If you have a minute, an honest review helps other people find us: [link]. Good or bad, we read all of them."
4. Don't ask on Yelp. Let Yelp reviews come on their own.
5. Answer every review, bad ones first, within a few days. A calm reply that says what you changed reads better to the next customer than a perfect score.
6. On your own site, show testimonials only from real customers who agreed in writing, keep the record, and don't hide the lukewarm ones. If a customer got anything for free, say so next to their words.
7. Staff: tell them in writing not to review the business, or to say they work there if they do. The FTC Q&A expects managers to act on red flags, such as a review from someone with an employee's surname.
A one-afternoon clean-up
If you have used an agency, bought a package or had staff "help" in the past, fix it before a platform or a letter does:
- List every place you have reviews: Google, Yelp, Facebook, Trustpilot, app stores, your own site.
- Take down testimonials on your site that you can't trace to a real customer who agreed to them, including any an AI tool wrote.
- Ask staff and family to delete their reviews or add a disclosure.
- Read your reputation agency's contract and ask in writing where its reviews come from. Under the rule, buying reviews you should have known were fake is enough.
- Turn off any review widget setting that holds low ratings for approval. That setting is how Fashion Nova got caught.
- Write the policy down in a few lines: how you ask, what you never reward, how staff disclose. Our one-page AI policy is a model for keeping it short.
If you are setting up a business listing at the same time, our guide to setting up a small business online in an afternoon covers the Google profile and the QR codes.
Put your review and testimonial policy on a public page that customers and staff can read: how you ask for reviews, what you never reward, and how insiders disclose. policies.biz59.net hosts policy pages with versions and a print view.
Write the policy pageSources
FTC, Federal Trade Commission announces final rule banning fake reviews and testimonials (14 Aug 2024; 5-0 vote; six prohibitions; AI-generated fake reviews named); FTC, Consumer Reviews and Testimonials Rule: Questions and Answers (in force 21 Oct 2024; hosting, generalized solicitation, AI avatars, gating, staff red flags); eCFR, 16 CFR 1.98 ($53,088 per violation for penalties assessed after 17 Jan 2025); Covington, FTC issues warning letters for violations of Consumer Reviews Rule (ten letters, 22 Dec 2025); Covington, FTC sets aside Rytr final order (Dec 2025); FTC, Final order against Rytr (Dec 2024); Fox LA, Fashion Nova to pay $4.2M to settle allegations it blocked negative online reviews (Jan 2022); Search Engine Roundtable, Google Maps blocked 292 million reviews (Apr 2026, with the 2024 figure) and Google, New ways we're protecting businesses on Maps; IANS, Google blocks over 170 mn policy-violating reviews in 2023; Trustpilot, Trust Report 2025 (4.5 million, 7.4%, 90% automatic) and Trustpilot, Trust (7.8 million in 2025); Google Maps user-contributed content policy (incentives, selective solicitation, conflict of interest); Yelp, Don't ask for reviews; Travers Smith, Fake or misleading consumer reviews and Cripps, Fake reviews (UK ban from 6 Apr 2025; CMA fines). All read October 2026.