AI generates fake reviews faster than platforms can catch them. Competitors buy negative reviews as weapons. And aggressive moderation is deleting legitimate feedback in the crossfire. For brands selling through marketplaces, the fake review economy has become a trust crisis with real casualties.
Online reviews have become some of the most valuable real estate in digital commerce. A single star rating can shift demand by double digits, and the collective weight of customer feedback now shapes nearly every purchasing decision. Trustpilot found that reviews influence 89% of global e-commerce revenue, with roughly half of consumers ranking positive reviews among their top purchasing factors. That concentration of influence has attracted exactly what it shouldn’t.
Generative AI now produces convincing testimonials at volumes that overwhelm human moderation, competitors weaponize negative reviews, and platforms deploy aggressive automated enforcement that sometimes catches legitimate feedback in the crossfire. The result is an environment where roughly 30% of online reviews are estimated to be fake, and where the trust signal that reviews were designed to create has become the vulnerability that attackers exploit.
In Los Angeles, contractor Natalia Piper learned what that looks like. Her company’s 5-star Google rating plunged to 3.6 after she received a WhatsApp message from Pakistan warning that someone had ordered 20 negative reviews of her business. The reviews appeared within days, and she paid $250 trying to stop the extortion, then paid again, before realizing the cycle would never end. Her story is not unusual. Small business owners across the country report similar patterns: extortion demands arriving via WhatsApp or Telegram, ratings cratering overnight, and no clear path to recovery through platform channels.
These are not isolated incidents. They represent the operational reality of a trust signal worth billions in commercial influence, now functioning as infrastructure for fraud.
How AI generates fake reviews at scale
What changed is the production method. Writing fake reviews used to require either paying human reviewers or accepting the stilted language that made automated reviews easy to spot. Generative AI eliminated both constraints.
The growth has been rapid. The Transparency Company observed AI-generated reviews growing 80% month-over-month since mid-2023, and the acceleration shows across industries. A study by Originality.ai found that nearly a quarter of Zillow agent reviews in 2025 were likely AI-generated, up from less than 4% in 2019. DoubleVerify’s Fraud Lab identified similar patterns in app stores, where one streaming app analysis found that half of all reviews were fake, identifiable only through uniform syntax and reviewers whose activity existed solely within that specific ecosystem.
The sophistication has evolved accordingly. Recent research found that AI-generated fake reviews demonstrate higher comprehensibility than human-written fakes while exhibiting lower specificity and exaggeration. They read more naturally, avoid the telltale signs that detection systems were trained to catch, and scale in ways that human review farms never could.
Platforms have responded with force. Google blocked or removed 170 million policy-violating reviews in 2023 alone, and Tripadvisor’s removals increased 50% year over year. But none of it has reversed the trend, because the economics still favor the fraudsters. The FTC estimates that businesses buying fake reviews see a 1,900% return on investment, and a fraudulent extra star can boost sales by double digits in the first two weeks. The upside is immediate and measurable; the downside, for most, never arrives.
Fake negative reviews as competitive weapons
The fake review problem extends beyond inflated ratings into something more destructive: a competitive weapon with a documented market for negative reviews targeting rivals.
Amazon’s October 2025 lawsuit against fake review brokers revealed the full service menu. The defendants operated dozens of websites offering fake five-star reviews, fake negatives aimed at rivals, fake seller feedback, and fake BBB business reviews. One site, ReviewServiceUSA.com, charged $50 per review. The court’s ruling ordered the transfer of all related domains to Amazon in the company’s most extensive website seizure to date.
On Amazon seller forums, merchants describe coordinated attacks that follow predictable patterns: clusters of one-star reviews appearing within days, written in similar language, sometimes mentioning competitor products by name. One seller shared a forwarded message from a customer showing a competitor’s offer: $10 for a negative written review, $20 to add a photo, another $20 for video. Another reported that a competitor’s CEO had left negative reviews using an identifiable seller account. The patterns are difficult to prove through platform channels and expensive to fight through legal ones.
Fashion Nova’s $4.2 million FTC settlement in 2022 showed the other side of review manipulation. The fast-fashion retailer used a third-party interface to automatically post four and five-star reviews while holding lower-starred reviews for approval that never came. For nearly four years, the company suppressed hundreds of thousands of negative reviews, artificially inflating product ratings until the FTC intervened in what became the agency’s first enforcement action of its kind.
The regulatory response is escalating. The FTC’s Consumer Review Rule, effective October 2024, enables civil penalties exceeding $50,000 per violation. In December 2025, the agency sent warning letters alerting companies to potential violations. The rule covers AI-generated fakes, purchasing negative reviews to harm competitors, and suppressing legitimate negative feedback.
When fake review moderation backfires
Google’s aggressive response to the fake review crisis has created a different problem: legitimate reviews disappearing alongside fraudulent ones.
Between January and July 2025, Google’s review deletion rates increased by more than 600%. But the deletions were not targeted at obvious fakes: analysis revealed that more than a third of deleted reviews carried five-star ratings, and businesses reported losing genuine positive reviews accumulated over years. One restaurant location saw 76 reviews deleted, spanning four years of feedback across all star ratings.
The pattern suggests that AI-powered moderation, deployed at scale, is generating substantial false positives. For businesses competing in local search, this creates an asymmetry that rewards past bad actors: companies that played by the rules lose authentic social proof, while competitors who purchased fake reviews before the crackdown may retain inflated ratings if their fraudulent reviews don’t match current detection patterns.
Consumer behavior reflects the uncertainty. Confidence in spotting fakes has improved, but that awareness translates directly into lost sales: over half of consumers will not purchase a product if they suspect fake reviews, and the suspicion doesn’t distinguish between genuine and fraudulent listings. The trust signal is degrading for everyone, honest businesses and manipulative ones alike.
Why fake reviews are a brand protection crisis
For companies selling through marketplaces, fake reviews create brand impersonation exposure that traditional monitoring doesn’t catch.
The damage flows in multiple directions. Counterfeit sellers use fake positive reviews to move fraudulent products, leading customers to associate poor quality with the legitimate brand. Competitors deploy fake negatives that tank algorithmic visibility, while platforms responding with aggressive enforcement sometimes delete the authentic testimonials brands spent years building. Research shows that 63% of consumers blame the legitimate brand when they’re victimized by impersonation, regardless of involvement.
Defensive measures exist but require resources most merchants lack, and traditional security tools weren’t built for this kind of monitoring. The FTC’s enforcement actions and Amazon’s lawsuits signal that regulatory pressure is increasing, but for now, the burden falls on brands to prove manipulation rather than on platforms to prevent it.
