In digital markets, the failure to adapt market definition tools risks undermining effective competition enforcement. In this open access contribution to the Journal of European Competition Law & Practice, Fabian Ziermann argues that defining markets in AI and video gaming must center on dataset substitutability. Using analogies from medicine, where cancer models cannot be trained on diabetes data, and gaming, where AI trained on one game mode fails in another, he illustrates why data markets should be delineated by dataset, not product.
Through case examples from Amazon, OpenAI, and Fortnite, he shows how this approach can address entrenched power in data-driven sectors. He further highlights that video games have long embodied AI-like structures, predating Big Tech’s dominance, and that competition authorities should treat AI and gaming as joint enforcement priorities. The article offers a concrete path forward for regulators seeking precision in defining digital markets shaped by algorithmic logic and data dependency, based on Amazon Marketplace. His contribution can be accessed here.