Amazon is carrying out an update to its measures to combat fake reviews, including the integration of Artificial Intelligence and machine learning models so that these ratings cannot be violated and cause sales that they should not.
The review system in electronic commerce is one of the ways that consumers have to launch into the purchase of an item that we may have never even seen in person. However, if this system is violated with false and/or erroneous reviews, the ratings lose their effectiveness and online trust in these online stores is reduced.
Amazon’s problem with reviews
Anyone who buys regularly at Amazon knows that, as in any other store, both online and physical, they can give you a pig in a poke. We tend to trust more products whose reviews have thousands of opinions and the sum of these approaches the score to five stars. So many people can’t be wrong, right?
However, a few months ago a case exploded in which fake SSDs with a similar appearance to the Samsung T5 or T7 were being sold. Inside, instead of a solid-state drive and its advantages, there were hacked memory cards inside a casing to fool buyers.
Victims of this scam wasted their money thinking they were buying a high-capacity hard drive with an excellent rating of almost 5 stars and over 8,000 reviews, while also putting their data integrity at risk by using these devices. The newspaper La Vanguarda, upon discovering the deception, alerted the company itself, as well as the Ministry of Consumption, which can investigate these frauds ex officio.
To the question about how this article was able to get so many positive reviews, because not everyone was able to sit idly by without discovering the hoax, the answer leads us to a certainty. There are companies that buy positive product reviews. This market is carried through Telegram groups. Users get paid to flood product listings with positive comments that bait potential buyers.
How to end these inflated scores
amazon has launch a new plan to combat fake reviews through public-private collaboration. Last year, 125 million customers contributed almost 1.5 billion reviews, so this system cannot be left unchecked and contaminated.
“Our goal is to ensure that every review posted on Amazon is authoritative and accurately reflects actual customer experiences. For this reason, at Amazon we encourage the publication of authentic reviews and categorically prohibit those that intentionally try to deceive customers, providing partial, inaccurate or unrelated information with the product or service reviewed.
To strengthen this area, a large amount of resources is being invested in the aim of proactively stopping fake reviews on Amazon, including the use of machine learning models that analyze a multitude of data to detect possible indicators of risk such as the use of related accounts, suspicious login activities and review history, and other signs of unusual behavior. “In addition, our team of expert investigators use sophisticated fraud detection tools to analyze and prevent fake reviews from being published in our stores.”
The newspaper that denounced Amazon’s practices in the case of the hard drive celebrates that the objective is being set on buying positive votes, but points to another credibility leak. “The system of manipulation of evaluations that we denounce is based on the fact that one product receives the scores of another. Case in point: the scores that led one of these fake hard drives to be one of Amazon’s best-selling electronics products came from a very different product, a box of crayons.”
