Abstract
This paper examines whether competition law enforcement can remain effective under different AI development scenarios over the coming years. Economic and political power has become increasingly concentrated into a few AI companies, such as Big Tech. The growth of generative AI could further reinforce this concentration of power in Big Tech. The market power of these companies, and increasingly their involvement in AI, is a major focus for regulators such as the European Commission. Recent EU antitrust fines on Google alone run in the billions. The dynamism of technology markets such as AI can make it difficult for regulators to take effective action. If AI continues to develop rapidly over the coming years, propelled by the proliferation of generative AI, this ability to effectively enforce antitrust law may be further challenged. To help ensure regulators remain effective, EU competition law has been bolstered by a new tech-tailored, ex ante competition regime. These are likely to be critical tools to shape the market power of Big Tech but are largely untested. Exploring how these regulatory tools can be most effective in governing future AI development is a timely question for regulators, lawyers, companies, and citizens. This paper examines this question by considering the ‘effective enforceability’ of EU competition law and the Digital Markets Act under different AI development scenarios. By ‘effective enforceability’ of EU competition law we mean how well it achieves its policy objectives. We consider four factors: jurisdictional scope, potential loopholes, effectiveness of detection, and ability to remedy/sanction breaches. However, there is significant uncertainty as to how AI will develop in the coming years. Considering this, we propose an analytical framework based on five variables: key inputs, speed of development, AI capability, number of actors, and the nature/relationship of actors. In some of these scenarios, we argue EU competition law would struggle to address the power of the largest AI companies; but in many other scenarios it remains a powerful tool. This is a critical juncture for competition regulators. They stand at the dawn of emerging challenges presented by generative AI. With this paper, we hope to contribute to anticipatory governance at this important intersection of legal governance and technology.
Effective and future-proof competition law enforcement is crucial to ensuring this potentially transformative technology has widely distributed benefits, rather than concentrating power in a few hands.
