Amazon is using artificial intelligence (AI) to crack down on fake reviews and identify comments that are not genuine.
According to the BBC:
The tech giant has been grappling with fake review “brokers”, which are a huge problem for its shopping site.
Amazon has invested in machine learning models that analyse thousands of data points to help it detect the fraudulent behaviour.
But UK consumer group Which? says the action is still “nowhere near enough”.
Fake review brokers use third-party platforms, including social media and encrypted messaging services, to buy, sell and host fake reviews.
Fake reviews can sway customers to make purchasing decisions, for example over which laptop or children’s toy to buy, based on what they believe is genuine feedback from other shoppers, when in reality someone has been paid to write a glowing review to boost a seller’s ratings, or to undermine a rival firm.
They aren’t always easy to spot, although generic information, or a very high percentage of five star reviews can be a give-away.
In 2022, Amazon reported more than 23,000 social media groups, with over 46 million members and followers, that facilitated fake reviews.
Amazon has been using AI in the battle against fake reviews for several years, but the company says continued investment in more “sophisticated tools” should improve protection for customers and sellers on its platform.
The company said its fraud-detecting AI was able to look at a range of factors to calculate the likelihood that a review is fake. That can include the author’s relationship with other online accounts, their sign-in activity, review history, and any unusual behaviour.
“We use machine learning to look for suspicious accounts, to track the relationships between a purchasing account that’s leaving a review and someone selling that product,” Dharmesh Mehta, the head of Amazon’s customer trust team, told the BBC.
“Through a combination of both important vetting and really advanced machine learning and artificial intelligence – that’s looking at different signals or behaviours – we can stop those fake reviews before a customer ever encounters it,” he said.
