by University of Oxford
Pixabay/CC0 Public Domain is credited.
A study by computer scientists at Imperial College and the Oxford Internet Institute found that consumers are at danger of vendor collusion due to sophisticated computer algorithms used to establish prices in online marketplaces.
According to Dr. Luc Rocher, primary author of the paper “Adversarial Competition and Collusion in Algorithmic Markets,” “the public is at risk of ‘adversarial collusion’ due to the widespread use of intelligent algorithmics and dynamic pricing by online retailers.”
The study describes the dangers of “adversarial collusion,” an anti-competitive tactic wherein a strong corporation might influence other vendors who have inferior pricing algorithms. It was published in the journal Nature Machine Intelligence.
The study, which was co-authored by Professor Yves-Alexandre de Montjoye and Dr. Arnaud J. Tournier of Imperial, shows how “adversarial collusion” can have detrimental effects on customers by allowing participating businesses to coordinate their prices and possibly increase their profits at the expense of consumers.
Digital marketplaces are becoming more and more dominant in online business, according to Dr. Rocher. Thanks to digital technologies, retailers are now able to compete on various platforms around the world. Sellers employ advanced algorithms to establish pricing and react instantly to rivals’ prices in order to stay competitive.”
He goes on, “Our study demonstrates how clever machine learning algorithms may identify weaknesses in the system. Stronger algorithms may be manipulated by more advanced algorithms, which allows them to work together to raise prices for everyone.”
“Adversarial collusion raises new regulatory and enforcement questions,” says senior author Professor de Montjoye. Our findings highlight the necessity for regulatory bodies in the US, EU, and UK to take into account the subtle ways in which adversarial algorithmic pricing systems may damage consumers and erode online markets’ ability to compete.”
“We believe further study is needed with academics and policymakers working together to try to address these important issues for the wider benefit of society.”
The researchers believe that adversarial manipulations of algorithmic pricing should not be covered by present competition rules or enforcement priorities, and they call on policymakers and regulatory authorities to take this into consideration.