Everything Else

How to Detect Fake Reviews on E-commerce Sites

How to Detect Fake Reviews on E-commerce Sites

Reading Time: 3 minutes

Are humans biased?

How well do humans spot fake reviews on e-commerce platforms? Not well at all according to a recent study conducted by Cornell University. The study found that humans suffer from what is known as a “truth bias”. This means that we as humans assume that what we read is true until proven otherwise with substantial evidence.

On the other end of the spectrum, when humans are specifically trained at detecting deception, they can become overly skeptical. This means that their detections of fraud can occur too often, resulting in severe inaccuracies. So it’s evident that humans aren’t exactly the most reliable detectors of fraud as bias can influence both ends of this spectrum.

How accurate is A.I. at detecting fake reviews?

Cornell decided to use Artificial Intelligence (A.I.) to detect false reviews on hotel sites which resulted in a whopping 90% accuracy. Three key findings resulted from this experiment:

  1. Truthful reviews tend to contain concrete words specifically relating to the hotel in question, such as, “bathroom” and “bed”. However, fake reviews are more vague and opt for setting the scene, i.e. talking about themselves and their general experience during their stay.
  2. Deceivers tend to use more verbs, while honest reviews contain more nouns.
  3. Fake reviews include more first person pronouns in order to over-compensate their illegitimacy.

E-commerce customers can use these three key findings to make their own judgements when it comes to online reviews on e-commerce sites. However, what if this needed to be conducted on a much larger scale? For example, if an e-commerce site has an abundance of product listings, there will more than likely be plenty of reviews of those products. Is the use of human detectors a practical way for online marketplaces to find these fake reviews? Probably not. A.I. is much more efficient when it comes to cost, resources, and speed, and most importantly: accuracy.

Fake Reviews

Do fake reviews really matter?

Now that we have a stronger understanding of what constitutes a fake review, you may be thinking: do false reviews really make that much of an impact when it comes to returning customers? Well the short answer is: yes. In fact, 90% of shoppers’ buying decisions are influenced by online reviews. So fake reviews can be extremely damaging to a brand’s reputation as they completely betray any loyalty that brand worked to create between them and their customers.

Be Sure To Also Read ->Counterfeit Detection Use Case Overview

Is there a culture of skepticism in e-commerce?

What brands and e-commerce sites may not be aware of is how much weight reviews carry when it comes to trust among their customers. 95% of consumers suspect censorship or false reviews are at play when bad reviews are nowhere to be found online. Skepticism in e-commerce, much like almost every place on the internet, is inevitable to an extent. However, according to Quartz India56% of online shoppers don’t trust product reviews, 65% don’t trust product ratings, and a whopping 72% of Indian e-commerce buyers are weary of incorrect information and believe fake reviews have become the norm. Is this really the right direction for e-commerce to be moving in?

Online shoppers can take matters into their own hands and decide for themselves whether or not a review of a product is fake. However, it is primarily the e-commerce site’s responsibility to ensure that its customers do not have to do this. Of course, brands shouldn’t be writing false reviews of their products in the first place, but until there is a way to stop them, elimination will win out over prevention.

What is the solution?

So what can e-commerce sites and online marketplaces do to protect their users? Using A.I. to detect and eliminate fake reviews on their sites is the most accurate way for online marketplaces to combat the scourge of fraud and deception that are rife online. A good example would be e-commerce giant, Amazon, who has already taken matters into their own hands. They have adopted an A.I. machine learning system that they built in-house to boost the weight of verified and newer customer purchase reviews and reviews that are deemed helpful by other users. This is just one way Artificial Intelligence can detect and weed out fake reviews in online retail.

LogoGrab specializes in logo detection technology used by leading e-commerce platforms to protect the legitimate brands that sell products on their sites against counterfeiters. This is slightly different to the A.I. technology used by Amazon to detect fake reviews, but it is just another way in which Artificial Intelligence can be used to eliminate fraud in e-commerce.

So A.I. can be used across e-commerce platforms for both Fraud Prevention and Brand Protection. Click the link to learn more about our Counterfeit Detection Visual-AI solution.

Book A Demo


  • The visual AI People Video: We Are VISUA - Discover why we’re the Visual-AI people - delivering cutting-edge computer vision solutions ranging from logo/mark detection and counterfeit product detection to holographic authentication and phishing detection... Posted in: Everything Else
  • shopping on phone 4 Ecommerce Trends – Updated for 2022 - Reading Time: 3 minutes As 2022 is well underway, new trends are beginning to emerge within the world of ecommerce. The online retail industry saw an […] Posted in: Everything Else
  • Visual-AI Logo detection pepsi-cola logo 8 Visual-AI FAQs - Reading Time: 5 minutes 1. What is Visual-AI? Visual-AI is the identification of objects and logos within an image. When using Visual-AI technology, images or videos […] Posted in: Everything Else
Could Computer Vision be the Solution to the International Luggage Issue?

Reading Time: 4 minutes The explosion in world travel after the easing of lockdowns and travel restrictions has seen airports struggle to cope. One of the […]

Everything Else Technology
Franco De Bonis Interview With Private Internet Access

Reading Time: < 1 minute VISUA Marketing Director Franco De Bonis sat down to chat with Private Internet Access recently to discuss our offering. Franco details the […]

Everything Else
The History of the Logo

Reading Time: 4 minutes The beginning of logo design We all know that a logo has the potential to be a very powerful asset for a […]

Everything Else

Trusted by the world's leading platforms, marketplaces and agencies

Integrate Visual-AI Into Your Platform

Seamlessly integrating our API is quick and easy, and if you have questions, there are real people here to help. So start today; complete the contact form and our team will get straight back to you.

  • This field is for validation purposes and should be left unchanged.