Amazon Taps Generative AI for Product Review Summaries
Amazon has announced new AI-generated customer review highlights to help customers shop confidently and find their needed products.
The AI-powered feature gives users a short paragraph on the product detail page and showcases product features. Customer sentiment shared across written reviews help customers quickly determine if the product is right for them. The AI-generated review highlights allow customers to easily pull reviews that mention specific product attributes and feature key product insights.
The feature is now available to a subset of mobile shoppers in the U.S. and has been incorporated into a broad range of products within Amazon.
Amazon will review customer feedback when considering future innovations, possibly expanding the review highlight features to include additional categories and customers in the coming months.
Amazon will also address fake reviews within their community guidelines to help both their machine learning and human moderators keep the community safe and the reviews authentic.
A study conducted by Fakespot in 2020 revealed that 30.9% of online customer reviews were deemed fraudulent. The report looked at reviews from online retailers like Amazon, BestBuy.com, eBay, Sephora.com, and Walmart.com, along with two million Shopify e-commerce websites. The data at that time showed that fake reviews were trending upward and often hit all-time highs during the holiday shopping season, including Black Friday through Cyber Monday.
Through growing investments, Amazon uses machine learning modules to analyze thousands of data points to help detect risk, including sign-in activity, review history, relations to other accounts, and anything that could be a red flag for unusual behavior. Amazon also uses expert investigators using sophisticated fraud-detection tools to analyze and stop fake reviews on the site.
The new AI-generated technology only uses a trusted review corpus from verified purchases, making it easy for customers to understand the community’s opinion quickly. These efforts allowed Amazon to block over 200 million suspected fake reviews in 2022.
In June 2023, Amazon filed four new lawsuits against fraudsters using fake reviews to mislead Amazon customers. The “fake review broker” industry approaches customers directly and solicits them to write fake reviews in exchange for complimentary products, money, or other incentives.
Amazon began sharing customer reviews on its website in 1995, producing continued updates to the experience to make it more beneficial to those leaving reviews and those making purchases. In 2019, Amazon created a quick star rating system, allowing users to leave reviews without writing a full-text review.
In 2022 alone, 125 million customers contributed about 1.5 billion reviews and ratings, leaving 45 reviews every second within Amazon stores.
