How Amazon is using AI to remove fake customer reviews

Amazon is working on various methods to ensure genuine customer reviews, but there have also been reports in the media about how to spot fake reviews on the platform. With the boom in artificial intelligence over the past year, several companies have introduced AI-powered solutions to enable artificial intelligence easily and effectively. Amazon uses AI to eliminate fake reviews.

Amazon says that when a customer submits a review before being published online, the company`s AI solution analyses the review for known indicators that the review is fake. While a vast majority of reviews pass Amazon`s high bar for authenticity and get posted right away, some go through the company`s scrutiny.

“If Amazon is confident the review is fake, they move quickly to block or remove the review and take further action when necessary, including revoking a customer`s review permissions, blocking bad actor accounts, and even litigating against the parties involved,” the company said.

The AI part

Amazon says that its machine learning models analyze proprietary data, including whether the seller has invested in ads (which may be driving additional reviews), customer-submitted reports of abuse, risky behavioral patterns, review history, and more.

These large language models (LLMs) and natural language processing techniques work in tandem to analyze anomalies in data that might indicate that a review is fake or incentivized with a gift card, free product, or some other form of reimbursement.

The company also notes that it uses deep graph neural networks to analyze and understand complex relationships and behavioral patterns to detect and eliminate groups of malicious actors.

“The difference between real and fake reviews is not always obvious to people outside of Amazon,” said Josh Meek, senior data science manager for Amazon’s Fraud Abuse and Prevention team.

“For example, a product may quickly gather reviews because the seller invests in advertising or offers a good product at a good price. Or because it contains bad grammar. Customers may think the reviews are fake,” Meek added.

Amazon also calls in expert investigators when a review is questionable or additional evidence is needed. In 2022, Amazon blocked more than 200 million suspected fake reviews in stores around the world.

Leave a comment