Fake Review Statistics 2026: What the Research Actually Says

Fake review statistics get copied from article to article until nobody remembers where they came from. This page lists only figures we could trace to the organization that produced them, with the method and the main caveat for each. We update it when new research is published.

Key Fake Review Statistics at a Glance

How Many Amazon Reviews Are Fake?

Nobody outside Amazon knows exactly, and Amazon doesn't publish a percentage. Three sources give a reasonable range:

• Independent research: the UK Department for Business and Trade study estimated 11–15% of reviews were fake in three large categories across nine platforms (the platforms were not named).
• Fakespot's model (vendor): before it closed, Fakespot reported that about 42% of 720 million Amazon reviews it assessed from March to September 2020 were unreliable, up from about 36% a year earlier (Bloomberg via Digital Commerce 360). "Unreliable" is broader than "fake," so this is an upper-bound style measure.
• Enforcement volume: Amazon blocked more than 275 million suspected fake reviews in 2024. That tells you the scale of attempts, not the share that slips through.

How Common Are AI-Generated Reviews?

Two AI-detection companies have published Amazon studies. Both sell detectors, and neither reported independent accuracy testing on review text, so read these as indicators:

• Pangram Labs (May 2026): about 3% of ~30,000 front-page reviews on 500 bestsellers were AI-generated with high certainty. 74% of the AI reviews gave 5 stars versus 59% of human reviews; 10% gave 1 star versus 22%.
• Originality.ai (November 2025): AI-heavy reviews rose about 400% after ChatGPT's launch. Extreme ratings (1 or 5 stars) were about 1.3× more likely to be flagged, and verified reviews about 1.4× less likely.

The two disagree on Verified Purchase: Pangram found 93% of AI reviews were verified, while Originality.ai found verified reviews less likely to be AI. Different samples (front-page bestsellers versus random reviews) likely explain part of the gap.

How Fake Reviews Affect What People Buy

The UK study ran an experiment with real shoppers. Clumsy fakes backfired: consumers were 5.3% less likely to buy products with poorly written fake reviews. Subtle fakes worked: they raised purchase likelihood by 3.1%, and by 9.2% for products over £80. A banner saying the platform moderates reviews had no significant effect.

The academic study of US Amazon sellers found the boost is temporary. After sellers stop buying reviews, ratings fall and one-star reviews rise, particularly for newer products (He et al.).

Fake Review Laws and Penalties

• United States: the FTC's rule on fake reviews and testimonials took effect on October 21, 2024 and covers AI-generated reviews. Maximum civil penalty: $53,088 per violation (the 2025 amount, which the FTC kept for 2026). See our FTC rule explainer.
• United Kingdom: the DMCC Act banned fake and concealed incentivized reviews from April 6, 2025, with fines of up to 10% of global turnover (CMS).

Numbers We Don't Use

You'll see some much larger figures online, such as unsourced claims that "one in three" reviews is fake or that fake reviews cost consumers hundreds of billions of dollars a year. We couldn't trace those to published research with a stated method, so they aren't on this page. (Fakespot's 30.9% figure above is sourced, but it comes from one company's model across the marketplaces it covered.) If you have a primary source for a figure we've missed, send it to us.

Citing This Page

You're welcome to cite these statistics. Please link to the original source for each figure and to this page as the compilation: FakeFind, "Fake Review Statistics 2026," fakefind.ai/blog/fake-review-statistics.