Algorithms and artificial intelligence can facilitate price-fixing and companies that use programmatic tools may need to start training their AI to steer clear of cartel enforcement action by the US Department of Justice, antitrust division chief Jonathan Kanter said last week.
As a growing number of companies embrace the emerging technology of AI, regulators, scholars, and technology developers alikehave grown worried that their automated systems may unfairly discriminate against protected classes of individuals, risking legal, regulatory, and reputational trouble. Fear of algorithmic bias has slowed deployment of AI in many sectors.
Market realities spurred by technology are relevant in a cartel case, Kanter said, signalling what is likely to be a close watch by the DOJ over the effects of these technologies as they become more prevalent. Kanter also discussed the division’s updates to its leniency policy, expanding the regulator’s litigation capabilities and improving cooperation with domestic and international enforcement partners.
Speaking before the International Competition Network’s annual conference, Kanter also highlighted the changes brought by the recent COVID-19 emergency. “The pandemic compelled us to find creative ways to maintain and deepen our international cooperation efforts,” said Assistant Attorney General Kanter. “Nevertheless, it is wonderful to engage again in-person with our ICN counterparts, especially to discuss the pressing competition issues we currently face.”
The conference showcased the achievements of the ICN’s Advocacy, Agency Effectiveness, Cartel, Merger and Unilateral Conduct Working Groups and examined a range of competition enforcement and policy issues. A main theme of the conference focused on planning for the third decade of the network. The continuing impact of the COVID-19 pandemic and digital markets on competition law also featured prominently
Want more news? Subscribe to CPI’s free daily newsletter for more headlines and updates on antitrust developments around the world.