The EU AI Act (Regulation (EU) 2024/1689) represents a significant departure from the EU’s traditionally restrained regulatory approach to commercial arbitration. The Act classifies certain use cases of AI in arbitration as potentially “high-risk” and introduces stringent compliance obligations for legal tech providers, arbitral institutions and arbitrators. This article argues that the Act’s application to arbitration disrupts the long-standing balance between party autonomy, procedural flexibility and regulatory oversight that has characterised the EU’s treatment of the field. It also highlights the challenges of reconciling its rigid framework with key aspects of arbitration – namely, party autonomy, confidentiality and procedural flexibility. The article concludes by proposing a full or partial carve-out of commercial arbitration from the scope of the AI Act’s high-risk provisions.

Artificial Intelligence (AI) has started to impact many facets of the economy and people’s routine activities. This article contributes to our understanding of how the legal system is reacting to the ongoing uptake of AI and the disputes or right infringements this uptake creates. Select legal cases regarding the use of AI technology for automated decisions are reviewed, with a focus on filings in Europe and the USA. This exercise reveals which type of legal challenges can be expected when it comes to deploying automated systems in these jurisdictions. Additionally, incipient regulatory efforts targeting AI on both sides of the North Atlantic are introduced and briefly discussed. The paper sheds light on how different legal systems accommodate an emerging technology with disruptive potential and offers a mapping of exemplary legal risks for prospective actors or organisations seeking to develop and deploy AI.