If you run a restaurant or dental practice, you’ve probably been pitched—or at least tempted by—services that promise to “boost your reviews.” Some are legitimate request campaigns. Others use generative AI to write reviews that look real. The second category is exploding, and platforms are building detection systems in response.
This matters because even if you never touch fake reviews, you need to know the signals platforms watch, the risks to your profile if you’re flagged, and how to protect a legitimately earned review stream from being swept up in a crackdown.
Here’s what we’re seeing, what the platforms have said publicly, and what it means for owner-operators who want to stay clean.
Why AI-Written Reviews Are Surging
Generative text models became cheap and accessible in late 2022. By mid-2023, offshore services and grey-market agencies started offering “managed review growth” at scale. The pitch is seductive: five-star reviews written in fluent English, posted from real-looking accounts, for a few dollars each.
The volume has grown enough that Google, Yelp, and Trustpilot have all updated their content policies and detection infrastructure in the past 18 months. Yelp’s Q3 2023 Trust & Safety report noted a “marked increase in organized efforts to post inauthentic content,” and Google’s spam-fighting teams have mentioned AI-generated text in multiple public updates.
For owner-operators, the risk isn’t just ethical. If your profile gets flagged—even for reviews you didn’t solicit—your visibility tanks, your existing reviews may be suppressed, and reinstatement is slow and uncertain.
Detection Signals Platforms Are Using
Platforms don’t publish their full playbook, but a combination of policy documents, patent filings, and observed enforcement patterns reveal the signals commonly tracked.
Linguistic Patterns
Early-generation AI reviews were formulaic: overly enthusiastic, repetitive sentence structures, generic praise (“the staff was amazing,” “highly recommend”). Current models are better, but they still leave fingerprints.
Platforms look for:
- Unusually uniform sentence length and complexity across multiple reviews in a short window
- Repetition of rare word combinations (AI models favor certain phrasings)
- Lack of specific, verifiable details—no dish names, no staff names, no temporal markers
- Over-indexing on abstract superlatives (“impeccable,” “unparalleled,” “phenomenal”) without concrete examples
If five reviews posted in the same week all use the phrase “attention to detail” but none mention your actual menu or a specific visit date, that’s a red flag.
Account Behavior
Even a perfectly written review gets flagged if it comes from a suspicious account. Platforms examine:
- Account age and review history (new accounts posting only five-star reviews are high-risk)
- Geographic clustering (ten “local guides” all reviewing the same three businesses in different states)
- Posting velocity (an account that goes from zero reviews to fifteen in a week)
- Device and IP patterns (reviews posted from the same device or VPN exit node)
This is why black-hat services now “age” accounts for months before selling them. It’s an arms race, and platforms are investing heavily on the detection side.
Temporal Clustering
A sudden spike in five-star reviews—especially reviews with similar length and tone—trips automated filters. Platforms expect review volume to track foot traffic and seasonality. If you’re a neighborhood dental practice and you go from two reviews a month to twenty in a week, the system assumes manipulation.
Legitimate campaigns (like a post-visit email reminder) do cause volume spikes, but they’re usually accompanied by variance in star ratings, review length, and content. A spike of uniform five-star reviews is almost always flagged for manual review.
Cross-Platform Correlation
Google, Yelp, and Facebook don’t share data directly, but they do compare patterns. If the same text (or near-duplicates) appears on multiple platforms under different account names, both copies get suppressed. AI services sometimes post identical reviews across platforms to “maximize ROI,” which makes detection trivial.
How Platforms Are Responding
Public enforcement has ramped up. Google’s transparency reports show millions of fake reviews removed each quarter, and Yelp’s Consumer Alerts program now flags businesses suspected of review fraud directly on their profile page—a scarlet letter that can persist even after cleanup.
In 2023, the Federal Trade Commission finalized a rule making it illegal to buy, sell, or post fake reviews. Penalties include fines up to $50,000 per violation. The rule explicitly covers AI-generated reviews, even if the business didn’t write them directly. Liability extends to the business that benefits.
Platforms are also testing generative classifiers—AI trained to detect AI. Google’s spam team has published research on using embeddings and anomaly detection to spot synthetic text. Yelp’s recommendation algorithm now includes a “review quality” score that weights authenticity signals.
For owner-operators, this means the window for “getting away with it” is closing fast. Even if you used a service two years ago and nothing happened, retroactive enforcement is common when platforms improve their models.
The Risk to Businesses
If your profile is flagged, the damage is immediate and hard to reverse.
Visibility Loss
Google and Yelp both suppress suspected fake reviews, but they may also downrank your entire profile in local search. A restaurant that drops from position two to position twelve in “near me” searches can lose half its discovery traffic overnight.
Consumer Alerts
Yelp’s Consumer Alert is the most visible punishment. A red banner appears at the top of your profile warning users that “someone associated with this business” may have tried to manipulate reviews. The alert stays live until Yelp’s trust team manually removes it—often months, sometimes longer.
Restaurants and dental practices report 20–40% drops in appointment requests or reservation clicks while an alert is active. The reputational damage lingers even after removal.
Legal Exposure
The FTC rule creates direct liability. If a competitor or customer reports you, the agency can investigate. Settlements are public, expensive, and permanent. Even if you argue you didn’t know the reviews were fake, the rule imputes knowledge if you hired the service.
How to Stay Clean
The safest path is also the slowest: earn reviews the boring way.
Request Reviews, Don’t Write Them
Ask happy customers to leave feedback. Make it easy with a short link or QR code. Train your front-of-house or front-desk team to mention it at checkout. This works—it’s just not instant. In our validation set of NYC businesses, the median restaurant earned about 15 reviews per month, and timing the request mattered more than the medium.
Don’t Incentivize Specific Star Ratings
You can offer a discount or entry into a drawing for leaving any review, but you cannot condition the reward on a five-star rating. That’s considered review gating, and both Google and Yelp prohibit it. For dental practices, compliance rules are even stricter depending on your state.
Respond to Reviews Yourself
Platforms look at response patterns, too. If your responses are all written in the same generative style as your incoming reviews, that’s a correlation flag. Write your own replies. Keep them specific. Mention details from the review. Show that a real human read it. Our analysis of overused owner phrases shows that even legitimate responses can sound canned if you lean too hard on templates.
Monitor Your Profile for Suspicious Activity
If reviews start appearing that you didn’t request—especially generic five-star posts from accounts with no history—report them. Don’t assume they’re helpful. They put your profile at risk, and platforms may assume you solicited them. Automated review monitoring helps you catch anomalies early, before they cluster into a pattern that triggers enforcement.
What Get Kandid Watches
The monthly Get Kandid Report tracks review volume, velocity, and sentiment trends. We flag unusual spikes, monitor review length and keyword patterns, and surface reviews that lack verifiable details. If you see a sudden influx of reviews that all sound similar, you’ll know within days—not months later when Google sends a policy warning.
The Starter plan ($29/month, annual saves 20%) includes the core Report. Pro and Business tiers ($59 and $99) add the Competitor Report and faster Email Alerts for negative reviews, so you can respond before a one-star review sits unanswered for a week. We draft suggested replies, you copy and paste—we never post on your behalf.
The first report is free, no card required, no sales call. Request your sample report to see what we surface.
Frequently Asked Questions
Can platforms really detect AI-written reviews if the text is good?
Yes, commonly. Even sophisticated AI text leaves patterns—uniform sentence structure, over-reliance on certain phrases, lack of specific temporal or sensory details. Platforms also weight account behavior and posting velocity, so even a perfectly written review from a suspicious account gets flagged. Detection is improving faster than evasion techniques.
What happens if a competitor posts fake reviews on my profile?
Report them immediately through the platform’s flagging tool. Include any evidence you have (unusual posting times, generic language, new accounts with no history). Platforms investigate, but the process is slow. If the reviews are clearly violating policy and you can demonstrate a pattern, enforcement is more likely. Document everything in case you need to escalate or respond publicly.
Is it risky to use a service that “requests” reviews on my behalf?
It depends on how they do it. If they send a legitimate post-visit email or SMS with a link to Google or Yelp, that’s fine. If they post reviews themselves, or if they use templates that generate uniform responses, you’re at risk. Ask for sample campaigns, check if they ever access review platforms directly, and avoid anyone who guarantees a specific star rating or volume. Transparency is the test.
How long does it take for platforms to remove suspected fake reviews?
Removal timelines vary. Google’s automated filters can suppress reviews within days, but manual review requests often take weeks. Yelp’s recommendation algorithm runs periodically and may filter reviews in batches. If your profile gets a Consumer Alert, reinstatement requires manual approval from Yelp’s trust team, which can take months. The safest strategy is prevention, not remediation.
The Bottom Line
AI-written reviews are easy to buy, hard to sustain, and increasingly risky to your business. Platforms are investing millions in detection, regulators are watching, and the penalties—both algorithmic and legal—are real.
If you’ve never touched fake reviews, stay vigilant. Monitor your profile for suspicious activity, request reviews the right way, and keep your response style human and specific. If you’ve used a grey-market service in the past, document what you did, stop immediately, and consider proactively reporting the reviews to minimize exposure.
The businesses that win long-term are the ones that earn reviews slowly, respond authentically, and build a track record platforms trust. It’s not glamorous, but it’s the only strategy that doesn’t blow up.