Amazon launched a generative AI tool on its e-commerce marketplace that helps shoppers make better buying decisions. The feature can read product reviews and create a summary of the most important points, according to a blog post published on Aug. 14.
Amazon said the summary appears as a short paragraph of text on the product detail page, which highlights the product features and customer sentiment mentioned in reviews. Online shoppers can use this information to understand the common themes across reviews.
What can Amazon’s AI do?
According to Vaughn Schermerhorn, director of community shopping at Amazon, the new AI tool offers key information about a product, such as whether most customers found it easy to use. The feature also points out more specific product details like weight or dimensions.
For example, an AI-generated review summary for a smartphone could highlight that many buyers have a positive view of its long battery life. The tool is aimed at minimizing the time people spend on checking what other users think of a product before making a purchase.
Several clickable keywords can be found below each AI-generated review summary. Positive reviews can earn a product an ‘ease of use’ tag from Amazon. By clicking on the tag, users are able to see another AI summary explaining why shoppers think the product is easy to use.
“The AI-generated review highlights also feature key product insights and allow customers to more easily surface reviews that mention certain product attributes,” Schermerhorn wrote in the blog post.
Amazon first added support for online reviews in 1995, a few months after it launched its eponymous e-commerce marketplace. Last year, over 125 million people wrote 1.5 billion reviews about products listed on Amazon’s online stores – the equivalent of 45 reviews per second.
Dealing with fake reviews
Per the blog post, the new generative AI review feature is being rolled out first to a subset of U.S. shoppers on mobile across a “broad” selection of products. Amazon will make the tool available to more people and products later this year, based on the feedback it receives.
“We are always testing, learning, and fine-tuning our AI models to improve the customer experience,” Schermerhorn said.
Amazon revealed the summary reviews generated by its AI only come from reviews that are submitted by verified customers. The AI feature also uses fraud detection algorithms to identify and remove fake reviews.
According to Schermerhorn, the company’s algorithms “analyze thousands of data points to detect risk, including relations to other accounts, sign-in activity, review history, and other indications of unusual behavior.”
In 2021, Amazon said it had blocked 200 million fake reviews the year before, TechCrunch reports. The American firm has been working to combat fake reviews for years, using a variety of methods, including lawsuits and fines against sellers who buy fake reviews.
For example, last year, it “sued the admins from 10,000 Facebook groups who were engaged in fake review brokering,” the report says.
Amazon is planning to introduce more generative AI features for its online stores in future. In May, Bloomberg reported that the company is working on a ChatGPT-like product search tool to “enable users to not only search for merchandise but also compare items.”