Learning how do I spot ai generated fake reviews online is becoming an essential consumer skill, because fake reviews are now industrial in scale. The Competition and Markets Authority estimates that fake reviews influence billions in UK consumer spending every year, and AI has poured fuel on the fire — what once required a click-farm now requires a prompt. I nearly booked a “5-star, 2,000-review” hotel in Cornwall last year until the reviews started setting off alarm bells. Here’s how to spot them.

The good news: AI-generated reviews have tells. They’re getting subtler, but they still fail in predictable ways — because they’re written by systems optimising for plausibility, not by humans remembering an experience.

How do I spot AI-generated fake reviews online?

The language tells

  • Suspiciously generic praise: “This product exceeded my expectations in every way. The quality is outstanding and I highly recommend it to everyone.” Real reviews mention specifics — the squeaky third-floor stair, the waiter named Dave, the delivery driver who left the parcel with the neighbour. Vague enthusiasm with zero detail is the biggest red flag.
  • Over-formal or oddly structured: AI reviews often read like essays — introduction, three points, conclusion — where humans ramble. Excessive use of stilted transitional phrases, perfect grammar throughout, and balanced pros-and-cons lists on a toaster review should raise eyebrows.
  • Emotional flatness: real reviewers are specific about feelings (“I was fuming when…”, “my kids absolutely loved…”). AI praise tends to be enthusiastic but oddly bloodless.
  • Repetitive phrasing across reviews: if twelve reviews all mention “exceptional customer service” and “highly recommend”, you’re reading a template, not twelve experiences.

The pattern tells

  • Review velocity: 200 five-star reviews appearing in a week for a product that’s been listed for years is a bought-review spike. Tools like ReviewMeta and Fakespot visualise this.
  • Reviewer profiles: click through. If the reviewer has only ever reviewed products from one brand, or their reviews span wildly unrelated categories posted on the same day, it’s a paid or bot account.
  • The rating distribution: genuine products have a spread — mostly positive with a tail of complaints. All 5-stars with no 3-star reviews is statistically suspicious. Real life is messier.
  • Verified purchase status: on Amazon, filter to verified purchases. It’s not foolproof (sellers game this too), but it filters out the laziest fakes.

The tools that help

Free tools like Fakespot and ReviewMeta analyse review patterns and give products a trust grade — worth running before any significant purchase. For local businesses, read the 2–4 star reviews specifically; they’re where honest customers live. And cross-reference: if a restaurant has 500 glowing Google reviews but is empty on a Saturday night, trust your eyes.

This matters beyond shopping, too — fake reviews distort everything from tradespeople to holiday lets. The same critical eye helps with spotting AI-written news articles, since the tells overlap. And if you’re a business owner worried about fake reviews targeting you, our guide to getting honest customer feedback covers building a genuine review base that drowns out the noise.

What the law says

Since 2024, fake reviews have been explicitly tackled under UK consumer protection law — the Digital Markets, Competition and Consumers Act bans fake reviews and requires platforms to take reasonable steps against them. If you spot obvious fakes, report them to the platform; enforcement is slowly catching up with the problem. The Which? investigations into fake reviews are excellent if you want to see the tactics documented in detail.

For businesses: earning trust in an age of suspicion

Here’s the flip side nobody talks about: the fake-review epidemic is making consumers suspicious of genuine reviews too. Businesses with legitimate five-star reviews now face customers who assume they’re bought. This is the tragedy of the commons playing out in real time — the fakers poison the well for everyone. So how does an honest business build credible social proof when trust is scarce?

The answer is specificity and verifiability. Encourage reviewers to mention details — the specific product, the staff member’s name, the particular experience. Detailed reviews are both more persuasive and harder to fake. Respond to reviews personally (not with templates), including the critical ones — a thoughtful response to a 2-star review builds more trust than ten generic 5-stars. Use verified-purchase and verified-visit systems where available. And diversify your proof: video testimonials, case studies with real names and photos, and user-generated content on social media are all harder to fake than text reviews and more convincing to sceptical buyers.

Also, report fake reviews aggressively — both the ones targeting you (competitors do leave fake negative reviews; it’s more common than people think) and the obviously fake positive ones on competitors. Platforms are slowly improving enforcement, partly driven by the UK’s new fake-review laws, and every report helps train the system. The businesses that win the trust game in the next decade won’t be the ones with the most reviews — they’ll be the ones with the most believable reviews. Authenticity is becoming a competitive advantage precisely because it’s getting rarer.

A final thought on the bigger picture: fake reviews are a trust tax everyone pays. Every minute you spend verifying reviews, every genuine business doubted, every good product overlooked because the reviews looked suspicious — that’s the cost of fakery, distributed across the whole economy. The long-term fix isn’t just better detection; it’s restructuring the incentives. Verified-purchase systems, reviewer accountability, platform liability, and the UK’s new fake-review laws all push in this direction. Until then, your scepticism is both self-protection and a small contribution to the solution: every fake you spot, report, and refuse to trust makes the ecosystem slightly healthier.

Your practical takeaway

Before your next significant purchase: run the product through Fakespot or ReviewMeta, read the 3-star reviews, check reviewer profiles for the top half-dozen raves, and look at the review timeline for spikes. It takes five minutes and it works. The broader lesson: in an AI-saturated internet, the scarcest and most valuable signal is specificity — trust the review that mentions the squeaky stair over the one that praises “exceptional quality” in the abstract.