Gleaning Insight from Antitrust Cases Using Machine Learning

Abstract

The application of AI and Machine Learning (ML) techniques is becoming a primary issue of investigation in the legal and regulatory domains. Antitrust agencies are in the spotlight because they represent the first arm of government regulation in that they reach new markets before Congress has had time to draft a more specific regulatory scheme. A question the antitrust community is asking is whether antitrust agencies are equipped with the appropriate tools and powers to face today’s increasingly dynamic markets. Our study aims to tackle this question by building and testing an antitrust machine learning (AML) application based on an unsupervised approach, devoid of any human intervention. It shows how a relatively simple algorithm can, in an autonomous manner, discover underlying patterns from past antitrust cases by computing the similarity between these cases based on their measurable characteristics. Our results, achieved with simple algorithms, show much promise from the use of AI for antitrust applications. AI, in its current form, cannot replace antitrust agencies such as the FTC. Instead, it is a valuable tool that antitrust agencies can exploit for efficiency, with the potential to aid in preliminary screening, analysis of cases, or ultimate decision-making. Our contribution aims to pave the way for future AI applications in market regulation, starting with antitrust regulation. Government adoption of emerging technologies, such as AI, appears to be key for ensuring consumer welfare and market efficiency in the age of AI and big data.