Amazon Tweaks Review System To Crack Down On Fakes, Highlight The Most Helpful Entries

Amazon is taking its recent crackdown on fraudulent reviews a step further by implementing a system to ensure customers see the most authentic – and helpful – reviews of a product first.

CNET reports that Amazon’s new machine-learning platform aims to identify and showcase the best assessments by overhauling an algorithm that places value on reviews.

Under the new system, the e-commerce giant will assign more weight to recent reviews, those that have been written by a verified purchaser and those that are seen as more helpful by other customers.

Additionally, the overall rating for a product will now be based on the new system rather than simply being the average of all the stars reviewers bestow upon it.

“Amazon is enhancing the customer reviews system, adding a few changes we hope will help make product feedback even more useful to customers,” Julie Law, an Amazon spokeswoman, tells ABC News. “The system will continue to learn which reviews are most helpful to customers and improve the experience over time.”

For the time being, the new system – which started on Friday – won’t be noticeable to most customers, CNET reports. That’s because the company will start by gradually altering star ratings and top reviews on product pages. Over time, the system will learn which kinds of reviews are most helpful and make additional tweaks.

News of the new system comes just two months after Amazon filed its first lawsuit against businesses it claims sell fake reviews to third-party sellers.

The lawsuit against four websites – BuyAmazonReviews.com, BayReviews.net, BuyReviewsNow.com. and BuyAzonReviews.com and their operators – accuses accuses the companies of an array of illegal business practices that undermine customer trust and the integrity of the online retailer.

Amazon says in the lawsuit that it actively monitors its website for “false, misleading and inauthentic reviews.” However, in certain instances that practice hasn’t been sufficient, leading the company to determine legal action was necessary.

The online retailer is seeking unspecified damages and requests that the sites cease and desist activities related to the retailer, provide information on each fake review created and the accounts of the people who paid for them.

Amazon looks to improve customer-reviews system with machine learning [CNET]
How Amazon Is Using Artificial Intelligence to Crack Down on Fake Reviews [ABC News]