How Amazon is summarising product reviews using Artificial Intelligence

AI-generated summaries is part of the new Amazon shopping experience built on generative AI.

Amazon is using Artificial Intelligence (AI) to assist users in finding the perfect product. They are experimenting with using generative AI to generate summaries of product reviews.

Product reviews

With so many items and reviews, it can be helpful for Amazon shoppers if AI can summarise the reviews for them. With Amazon's new one-tap rating system, customers can express product feedback without having to write detailed reviews. The function gives users an overview of what customers like and hate about a product they bought.

Amazon's new AI feature

Amazon recently began testing a feature in its shopping app that utilizes AI to summarise customer evaluations of certain products. It includes a quick summary of what customers liked and disliked about the product and a disclaimer that the summary was "AI-generated from the text of customer reviews." Amazon acknowledged testing the function. The business withheld precise information regarding how it operates or what AI models are being applied to generate text and summarise reviews.

Amazon has long relied on AI and machine learning to provide shoppers with tailored adverts and personalized recommendations. However, the recent hype surrounding generative AI and chatbots such as OpenAI's ChatGPT has prompted companies like Amazon to increase their focus on the technology. Using AI to summarise reviews could be beneficial to customers. Amazon has millions of things on its online site, and a single item can have hundreds of reviews. Through its one-tap ratings system, which allows customers to offer input without writing a full review, the company has made it easier for shoppers to review products. Recently, some customers have begun using chatbots to create Amazon reviews.

Challenges

Amazon has long struggled with trust. Many organizations have used a combination of human moderators and automatic systems to screen out fake reviews on the site and other online marketplaces. Some customers have even used chatbots to submit reviews on their behalf.

According to Wieczorek's LinkedIn post, summaries of AI-generated product reviews could help direct buyers to more recent reviews. It's been known for a long time that older products, which have had more time to accumulate evaluations, have an unfair edge over fresher (perhaps superior) entrants, he added.

Competition

Google is also developing similar technology. They define it as "With generative AI in Search, we can help you understand the full picture when you're shopping, making even the most considered and complex purchase decisions faster and much easier." It is now in limited testing. When you search for a product, you'll get a list of essential factors to consider and products that match the bill. You will also receive product descriptions with relevant, up-to-date reviews, ratings, pricing, and product photos."

Amazon and Google are both working towards a similar purchasing future. Instead of the decades-old search bar that returns a grid of product results, prompting the shopper to click on individual results, read reviews, and compare prices, the new experience will return fewer but more relevant results and, at least on Amazon, will most likely display an AI-generated summary of reviews. It is about transforming mountains of data on buying behaviour and consumer preferences into a more user-friendly shopping experience.

Conclusion

Amazon is introducing the AI feature when the technology is gaining popularity. It is because the ChatGPT bot became popular, causing a race in technology. Google, Microsoft, and Meta have all hopped on the AI bandwagon. People are concerned that technology may eliminate jobs in the media, market research, and even the legal profession.

Furthermore, in recent years, Amazon has abandoned consumer experience innovation, replacing it with advertisements. Amazon is being forced to return to it by the generative AI race. Summaries of reviews are a modest experiment lost among the long list of features on the product page, but they hint at much more to come.