Without Any Prejudice? The Antitrust Implication of the AI Act

This research paper questions to what extent the AI Act allows for algorithmic regulation and decision-making to take shape in competition law. It first summarises the algorithmicising of competition law proceedings. It then discusses algorithmic decision-making under the AI Act. Annex III of the AI Act qualifies as high-risk AI systems intended to be used by law enforcement authorities in the course of detection, investigation and prosecution of criminal offences. As some EU Member States criminalised competition law, this would mean the AI Act might apply to AI systems used in competition proceedings. Competition law is, however, not qualified as criminal in all Member States. This paper therefore questions the opportunity to use the dichotomic relation between hard core and peripheral law developed by both the European Court of Justice and the European Court of Human Rights. This paper argues, however, that such solution is doomed to failure. A contextual interpretation of the Proposal prevents any application of the AI Act to competition law through the back-door of peripheral criminal law. The paper concludes by imagining the competition landscape after the AI Act. On countries wherein competition law is criminal, national competition authorities will have to take that Regulation into account when developing AI systems that detect, investigate and prosecute competition infringements. As not all EU Member States have criminalised competition law, this would result in an imbalance between domestic legal orders that goes against the prime ambition of the AI Act, i.e., harmonisation.