Algorithms have gained significant attention in recent years given technological advancements in AI and growing amounts of data. Algorithms can be used for a range of purposes. Firms increasingly use algorithms to set prices. Algorithms allow consumers to rapidly search for products and receive personalised recommendations. However, while algorithms provide many benefits, they can also reduce competition, and harm consumers.

In June 2023, the OECD held a roundtable to discuss:

• algorithmic theories of harm and example cases
• whether existing competition law is sufficient to address algorithmic theories of harm and potential remedies 
• how competition authorities can investigate algorithms

KEY FINDINGS

The discussion highlighted the various ways in which algorithms can harm consumers, both through co-ordinated conduct (such as algorithmic collusion) and unilateral conduct (such as algorithmic exclusionary and exploitative abuses). Algorithms can change the assumptions on which competition law is based and competition authorities should continue to consider to what extent the existing law may need to change to reflect this. Finally, the experts and delegates discussed the range of methods available to investigate algorithms, as well as the breadth of evidence that an authority could consider. The most relevant technique is usually case-specific. An authority will not always need to adopt sophisticated technical approaches, and simpler methods or evidence may be sufficient in some cases. There have still been relatively few relevant cases and authorities often face several practical challenges when investigating an algorithm. However, authorities are increasingly developing in-house technical knowledge to build their capacity to take on cases involving algorithms, which is important, as these types of cases are likely to become more common.