Amazon employs generative AI to summarize customer reviews

Amazon employs generative AI to summarize customer reviews

Amazon Utilizes AI to Summarize Customer Reviews

Amazon AI

Shopping on Amazon can sometimes be overwhelming when faced with hundreds or even thousands of customer reviews for a popular product. However, Amazon has now introduced a convenient solution to this problem with the official rollout of its new AI-driven product review option. With this feature, mobile shoppers in the US can now access a concise AI-generated summary that captures the opinions of numerous buyers.

The generative AI technology offered by Amazon delivers a short paragraph on a product’s detail page, highlighting its core features and summarizing the overall customer sentiment. Additionally, the AI-generated summary allows users to select specific attributes of a product, such as “ease of use” or “reliability,” to explore excerpts from reviews that mention those factors. By providing this customer-centric summary, Amazon aims to streamline the shopping experience and help shoppers make informed decisions.

Amazon initiated testing of this new feature in June and has now expanded its availability to a wider range of products. When the feature is available for a particular product, shoppers will notice a customer reviews section that appears above the actual reviews, containing a concise summary generated by AI, along with a notice that it is AI-generated from customer review texts. Furthermore, choosing an attribute below the paragraph reveals a summary focused on that specific factor, accompanied by brief snippets from various customer reviews.

Newegg AI

However, Amazon isn’t the only online retailer incorporating AI-generated summaries of customer reviews. Newegg, a fellow vendor, recently announced its use of ChatGPT to condense customer reviews into “Review Bytes,” which are short snippets accompanied by longer paragraphs created by AI. For now, this feature is only accessible on the desktop version of the Newegg website and limited to products with a minimum number of reviews.

While the introduction of AI-generated summaries may appear beneficial, it naturally raises questions regarding accuracy and the exclusion of fake or false reviews. Trusting the AI to accurately reflect the collective sentiment of numerous reviews is crucial. In response to this concern, Amazon mentions that it continuously tests and refines its AI models to enhance the overall experience. However, detecting and preventing fake reviews has long been a challenge for Amazon and its customers.

To tackle fake reviews, Amazon invests significantly in resources that proactively detect and prevent their publication. Machine learning models analyze various data points, including account relationships, sign-in activity, and review history to identify fraudulent reviews and unusual behavior. Moreover, investigators employ fraud-detection tools to analyze and prevent fake reviews from impacting the marketplace. To ensure the reliability of the AI summaries, Amazon incorporates only trusted reviews from verified purchases.

Amazon Lawsuit

In its fight against fake reviews, Amazon has also taken legal action against fake review brokers who attempt to manipulate customer opinions. These brokers contact customers through websites, social media, and messaging services, enticing them to write fake reviews in exchange for compensation or free products. In 2022 alone, Amazon successfully blocked over 200 million suspected fake reviews from appearing online.

In summary, Amazon’s AI-driven product review option brings convenience to shoppers by presenting them with concise summaries of customer opinions. While the use of AI to summarize reviews may raise concerns about accuracy and fake reviews, Amazon assures its customers that it continually refines its AI models and invests in resources to combat fraudulent activities. By incorporating only trusted reviews from verified purchases, Amazon aims to provide a reliable and informed shopping experience for its users.


Reference: