Cooperation between companies developing artificial intelligence (AI) can help them create AI systems that are safe, secure, and with broadly shared benefits. Researchers have proposed a range of cooperation strategies, ranging from redistributing “windfall” profits to assistance to address the harmful dynamics of a competitive race for technological superiority.
A critical tension arises, however, between cooperation and the goal of competition law, which is to protect the very process of competition between rival companies. Whilst these potential conflicts are significant, they are currently underexplored in the literature. This paper examines the relationship between proposed forms of AI cooperation and competition law, focusing on the competition law of the European Union (EU).
EU competition law governs the behavior of the world’s largest AI companies, though many are based abroad, especially in the US. Its jurisdiction can extend to any foreign company that is active in the EU. Scrutiny of US “Big Tech” is also an area of strategic focus for the European Commission (EC).
This paper seeks to reconcile the cooperative AI development and competition law. It examines fourteen forms of AI cooperation, both those that are applicable today and longer-term strategies that will apply when AI development is more advanced. Where we identify potential tensions with EU competition law, we suggest mitigation steps. Our aim is to ensure the long-term sustainability of these important safeguards to the responsible and beneficial development of AI.