Newsletter
Generative AI has quickly found its way into a lot of industries, thanks in part due to its summarization and conversational skills. Now, in a recent development, Amazon has started rolling out a new feature that leverages AI to summarize user reviews, providing users with a general idea of the pros and cons of a product while also saving valuable time.
According to Amazon, this feature underwent initial testing several months ago and, following positive feedback, is now being introduced to a select group of users across the United States. In terms of functionality, the AI-generated review summaries will operate in a manner akin to the “Critics Consensus” and “Audience Says” summaries on platforms such as Rotten Tomatoes. For instance, if a user is thinking about purchasing a new LG smart TV and is uncertain about the user feedback. They can use the AI-generated summary feature, which will highlight the positive and negative attributes of the products.
In an effort to further strengthen the AI summaries feature, Amazon is also implementing clickable tags that act as navigational aids, highlighting prominent themes and common keywords gathered from the pool of customer reviews.
Fake user reviews
While Amazon’s efforts to provide AI-generated review summaries are a step in the right direction, the company’s struggles with fake reviews could undermine the whole point. This is because if the customer reviews are fake and falsely promote the product, the AI summaries will also convey the same narrative. However, it is important to note that Amazon has acknowledged these concerns and stated that it exclusively relies on information derived from verified purchases. Moreover, in doubtful cases, Amazon says it will employ human investigators equipped with advanced fraud-detection tools to analyze and counteract instances of fabricated feedback.
“We continue to allocate substantial resources towards proactively combating fake reviews. This encompasses machine learning models scrutinizing thousands of data points to pinpoint potential risks, including account relationships, sign-in activities, review histories, and other indicators of unusual behaviour,” says Amazon.