What happens when a customer sends a review? Before its online publication, Amazon uses Artificial Intelligence (AI) to analyze the review by checking for the presence of known indicators that reveal its falsity. The vast majority of reviews pass this authenticity check and are immediately published.
Amazon, here’s how it handles suspicious reviews
Reviews are one of the main reasons why i Customers love shopping on Amazon since its inception in 1995. Amazon makes it easy for them to leave trustworthy reviews to help the purchasing decisions of millions of other customers around the world. At the same time, the company makes it difficult for bad actors to exploit customers’ trust in the Amazon shopping experience.
However, if you suspect potential wrongdoing, several possibilities open up. If the review is known to be false, Amazon promptly blocks or removes it and, if necessary, takes further action such as revoking a customer’s privilege to review, blocking the accounts of bad actors, or even taking legal action against the reviewers. related parties.
Be one review is suspect, but more evidence needs to be collected, Amazon’s expert investigators, specifically trained to detect misconduct, will continue the verification process by looking for more data before taking further action. In 2022, Amazon verified and blocked, thanks to proactive activities, over 200 million suspicious reviews on its sites worldwide.
Amazon uses AI
Along with other tools, Amazon uses the latest developments in AI to block hundreds of millions of suspicious online reviews, manipulated ratings, fake accounts, and other abuse before customers can view them. Machine Learning (ML) models analyze a vast amount of first-party data, including whether or not the selling partner advertises (which could lead to more reviews), abuse reports submitted by customers, risk behavioral patterns, review history and much more.
Large Language Models (LLM) and techniques are exploited Natural Language Processing to analyze anomalies in this data that may indicate that a review is false or incentivized, for example, through gift cards, freebies or other forms of reimbursement. Amazon also uses Deep Graph Neural Networks (GNN) to analyze and understand complex relationships and behavior patterns with the goal of identifying and blocking groups of bad actors.
It is on this point that some of our detractors are wrong about the detection of fake reviews: They are forced to make important assumptions without having access to data that highlights the presence of a pattern of abuse. Using cutting-edge technology and proprietary data, Amazon is able to more accurately identify fake reviews by going beyond superficial indicators of abuse to identify deeper interconnections between bad actors.