The short answer: fake reviews written by AI are banned everywhere. The FTC's rule on consumer reviews covers "AI-generated fake reviews" by name (FTC), and Google requires every review to reflect a genuine experience (Google). But Google Maps' policy has no rule against a real customer using AI to help word a review of their own visit. The line people quote, "we don't allow reviews that are primarily generated by... artificial intelligence," comes from Google's Shopping product review policy, not Google Maps.
So the questions that matter are who the review is from, whether it describes a real experience, and whether the business tried to shape what it says. This guide covers each one, how platforms actually catch fake reviews, and how to offer customers an AI draft without crossing the line.
Key Takeaways
- AI fake reviews are illegal in the US. The FTC's rule bans businesses from creating reviews that misrepresent the reviewer, their experience or whether they had one, and civil penalties are possible for knowing violations (16 CFR 465.2).
- Google Maps tests for a genuine experience, not for AI. Its policy says contributions "should reflect a genuine experience at a place or business."
- The "no AI reviews" sentence is about Google Shopping. It's in Google Merchant Center's Product Ratings policy (Google).
- Businesses can't steer the content. Since April 2026, Google says merchants shouldn't "request that specific content be included" in reviews, including content that identifies a staff member.
- Yelp and Tripadvisor are stricter. Yelp removes reviews written or revised with AI tools, and Tripadvisor bans businesses from suggesting wording.
- Platforms catch fakes mostly by behavior. AI detectors are unreliable, so Google looks for patterns like copied reviews and sudden rating spikes.
What's Clearly Banned
There's no ambiguity about these:
- Reviews invented by AI. Reviews from people who don't exist, or who never visited, are fake under Google's policy and the FTC's rule. When it announced the rule, the FTC said it covers reviews that misrepresent "that they are by someone who does not exist, such as AI-generated fake reviews, or who did not have actual experience with the business" (FTC).
- Businesses writing reviews for customers. Posting a review on someone's behalf, from their account or yours, misrepresents who wrote it. Tripadvisor names "writing reviews on behalf of users" as a violation (Tripadvisor).
- Paying for reviews, AI-written or not. Google bans incentives for any review, and the FTC's rule bans incentives conditioned on a positive review.
The consequences are real. Google can remove violating reviews and restrict a profile, for example by blocking new reviews for a set period, unpublishing existing ones, or showing a warning that fake reviews were removed (Google). The FTC can seek civil penalties for knowing violations of its rule.
What Google Maps Policy Actually Says
Google's content policy for Maps says contributions "should reflect a genuine experience at a place or business," and that "a review or rating should reflect an actual experience with a business, and be genuine and unbiased." Under fake engagement, it bans "content that is not based on a real experience or does not accurately represent the location or product in question" (Google).
The policy doesn't mention AI. The test is the experience, not the tool used to type it.
What it does say about businesses, in the rating manipulation section, matters more. Merchants may not:
- offer incentives for any review;
- "selectively solicit positive reviews from customers";
- "require or pressure users to leave ratings or write reviews while on the premises, nor ... request that specific content be included," including reviews "that include specific content, including content that identifies a staff member."
Merchants may ask for reviews "without offering incentives to do so or attempting to influence the rating or the contents of the review."
The Google Rule People Misquote
Many articles say Google bans AI-written reviews, quoting: "We don't allow reviews that are primarily generated by an automated program or artificial intelligence application." That sentence is real, but it's from the Product Ratings policies in Google Merchant Center, which govern product reviews that retailers send to Google Shopping (Google). It isn't part of the Google Maps policy for reviews of local businesses.
That doesn't make AI-generated local reviews fine. A review generated by AI with no real customer behind it is a fake review under the Maps policy anyway. It just means the "no AI" sentence doesn't apply to Google Maps.
What the FTC Says
The FTC's rule on consumer reviews has been in force since October 21, 2024. It bans businesses from writing, creating, buying or selling reviews that misrepresent who the reviewer is or what they experienced (16 CFR 465.2). It doesn't apply to ordinary consumers writing about their own experience (FTC).
The FTC's case against Rytr, an AI writing tool, shows where regulators draw the line. Its 2024 complaint said Rytr's review generator produced reviews with "specific, often material details that had no relation to the user's input," so they "almost certainly would be false" for the people who posted them (FTC). In December 2025 the FTC set that order aside, saying that "treating as categorically illegal a generative AI tool merely because of the possibility that someone might use it for fraud is inconsistent with our precedents and common sense" (FTC).
The takeaway: an AI writing tool isn't illegal in itself. The problem is a review that says things the reviewer never experienced. The FTC's guidance on testimonials makes the same point: a business "shouldn't provide the text for a testimonial without a reasonable basis to conclude that the text is truthful about the testimonialist's experience" (FTC).
What Other Review Sites Say
- Yelp says reviews written with third-party AI tools and chatbots, including ones used to draft or revise a review, are subject to removal (Yelp). Yelp also asks businesses not to request reviews at all.
- Tripadvisor bans businesses from "suggesting specific terms, wording, or paraphrasing to be included in reviews" (Tripadvisor).
- Trustpilot requires reviews to be based on a genuine experience and screens every review with automated detection before it's published (Trustpilot).
How Platforms Catch Fake Reviews
Plenty of people now use AI to tidy up what they write, so it's fair to ask how a platform could tell, or stop it.
Mostly, they can't tell from the words. AI text detectors are unreliable. OpenAI withdrew its own detector in 2023 "due to its low rate of accuracy": in its tests it caught only 26% of AI-written text and wrongly flagged human writing 9% of the time (OpenAI). Nobody knows how many reviews are written with AI. Detector-based studies give rough estimates; one put about 19% of recent G2 software reviews down as likely AI-generated (Originality.ai), but those numbers inherit the detectors' errors.
So platforms watch behavior. Google describes models that look at longer-term signals, "like if a reviewer leaves the same review on multiple businesses or if a business receives a sudden spike in 1 or 5-star reviews," along with whether an account has targeted other listings before (Google). Since April 2026, when Google sees a sudden spike in spam reviews, it removes them, pauses new reviews on the profile and shows a banner explaining why (Google). In 2025 it blocked or removed more than 292 million policy-violating reviews. Trustpilot says its detection "goes beyond the content of reviews to analyse a large range of behavioral data points."
Practitioners tracking Google's review-posting blocks report the same thing. In cases Sterling Sky followed in 2026, the common signals were reviews left inside the business, QR codes at the counter, and bursts of 10 or more reviews in a day, with blocks following weeks of sustained high volume rather than a single spike (Sterling Sky).
What this means for you: the reviews that get caught aren't the ones a customer tidied up with AI. They're the ones that look manufactured: copied text, sudden floods, reviews pushed out on the premises. Asking every customer after their visit, one at a time, keeps your reviews looking like what they are.
When a Real Customer Uses AI
Neither Google's Maps policy nor the FTC's rule stops a customer from using AI to help word a review of their own visit. What matters is the result:
- It describes their own, real experience. Nothing invented, nothing exaggerated.
- It's accurate. Google's tips for reviewers ask them to be "as accurate as possible" and to include "both the positive and negative aspects" of a visit (Google).
- They post it themselves, from their own account.
Generic AI phrasing ("an exceptional experience from start to finish") tells readers nothing. The most useful reviews are specific, whoever typed them.
When a Business Offers Customers an AI Draft
This is the grey area, because the business is now involved in the wording. No rule from Google or the FTC addresses it directly, so the safe approach follows the rules that do exist:
| Principle | Why |
|---|---|
| The draft's facts come only from what the customer told you | Google requires a genuine experience; the FTC's Rytr case was about details unrelated to the user's input |
| The draft matches the customer's own rating | Google bans attempts to influence the rating |
| The business adds nothing it wants mentioned: no staff names, keywords or claims the customer didn't bring up | Google bans requesting specific content, including staff names |
| The customer can edit or replace every word, or skip it | It has to be their review |
| The customer posts it themselves, from their own account | Posting for someone misrepresents who wrote it |
| Every customer is asked the same way, whatever they rated | Google bans selectively soliciting positive reviews |
| No reward of any kind for the review | Google, Yelp, Tripadvisor and Trustpilot ban incentives |
| No drafts for Yelp or Tripadvisor | Yelp removes AI-drafted reviews; Tripadvisor bans suggested wording |
How Spokk's AI Review Drafts Follow These Rules
Most customers who are happy to leave a review stall at the empty box. Spokk's AI review drafts are there for that moment: they turn what the customer already told you on your feedback form into a few sentences in their voice, so they're checking and posting rather than starting from scratch. It's still their review, about their visit, posted by them.
| The rule | How Spokk's drafts handle it |
|---|---|
| Reflect a genuine experience | A draft is only written after a real customer fills in your feedback form, and only from what they gave you: their rating, their comment, the services they picked, the team members they rated and their answers to your questions. The AI is told never to invent details, people, products or results. |
| Don't request specific content | Your business description and your AI notes are used as background and style only. They can't put a claim, service or name in a customer's review that the customer didn't mention. |
| Don't influence the rating | The draft follows the rating they gave. A 2-star visit gets an honest 2-star draft, not a softened one. |
| It's their review | They can edit every word of the draft before they copy it, or ignore it and write their own. |
| The reviewer posts it | They copy it and post it themselves, from their own Google account. Spokk never posts reviews, and there's no draft without a customer who filled in your form. |
| Yelp and Tripadvisor rules | The draft isn't copied when a customer taps a Yelp or Tripadvisor button. |
Everything is in your control:
- On or off. Turn drafts on or off for each feedback form.
- What goes in. Choose whether drafts can use the company name, the customer's written feedback, service names and staff names, and switch each custom question in or out on its own.
- The voice. Use AI Review Instructions to set the tone, list words to avoid and name claims the AI should never make.
- Their language. Drafts are written in the language the customer chose for the form.
One setting worth a look: staff names appear in a draft only when the customer rated that person 5 stars themselves. Because Google's April 2026 wording singles out content "that identifies a staff member," switching staff names off is the cautious choice.
Drafts also make a measurable difference to whether people follow through. In Spokk's data, 1.4x more customers click through to Google when the request includes an AI-written draft (November 2025 to June 2026; these are click-throughs, not counted reviews).
All-in-one growth platform
Turn every visit into more revenue with Spokk
One platform handles your entire post-visit customer lifecycle. More reviews, more repeat visits, more referrals. No juggling separate tools.
Google Reviews
AI-drafted reviews customers actually post
SMS Automation
Feedback, review, and referral sequences on autopilot
Loyalty Program
QR check-in rewards that bring customers back
Referral Program
Every customer gets a unique link to send friends
Feedback Collection
Private channel for every customer after every visit
Testimonials
Video and text testimonials collected automatically
FAQs
Are AI-generated Google reviews allowed?
Fake ones aren't, anywhere. A review invented by AI, or posted for someone who never visited, breaks Google's policy and the FTC's rule on fake reviews, which names AI-generated fake reviews specifically. Google Maps' policy has no separate rule about a real customer using AI to help word a review of their own visit. The test is whether the review reflects a genuine experience.
Does Google ban reviews written with AI?
Not on Google Maps by name. Google's Maps content policy requires that reviews reflect a genuine experience and bans fake engagement, but it doesn't mention AI. The often-quoted line "We don't allow reviews that are primarily generated by an automated program or artificial intelligence application" is from Google Merchant Center's Product Ratings policy, which covers product reviews in Google Shopping, not reviews of local businesses.
Is it OK to use ChatGPT to write a Google review?
If it's your own review of a real visit, Google's Maps policy doesn't forbid it. Read it before you post: it should say only what happened and what you actually think, with nothing added or exaggerated. Don't do it for Yelp, which removes reviews written, drafted or revised with AI tools.
Can Google tell if a review was written by AI?
Not reliably from the words alone. AI text detectors make too many mistakes, and OpenAI withdrew its own detector in 2023 for low accuracy. Google describes catching fake reviews mainly by behavior: the same review posted on several businesses, a sudden spike in 1-star or 5-star reviews, and accounts with a history of suspicious activity.
Is it illegal to use AI to write reviews?
Writing fake reviews is. The FTC's rule on consumer reviews, in force since October 2024, bans businesses from writing, creating or selling reviews that misrepresent that the reviewer exists, had an experience, or what that experience was, and the FTC said the rule covers AI-generated fake reviews. A real customer describing their own experience is a different case, but a business should never generate reviews itself.
Can a business give customers an AI draft of their review?
No rule from Google or the FTC addresses it directly, so it depends on how it's done. On Google, the draft's facts should come only from what the customer actually told you, the customer should be free to change or discard it, they should post it themselves, and the business shouldn't add content it wants included, such as staff names or keywords. Don't offer drafts for Yelp or Tripadvisor, whose rules bar them.
Will Google remove reviews written with AI?
Google removes reviews that break its policies, including fake engagement and rating manipulation, and in 2025 it blocked or removed more than 292 million of them. It hasn't said it removes genuine reviews for being worded with AI. Generic text that says nothing specific is still a bad idea: it helps no one and reads like spam.
