Why good friction matters in international arbitration in the age of AI

This article examines the growing role of artificial intelligence (AI) in international arbitration and argues that efficiency without engagement threatens the foundations of legal judgment. Drawing on cognitive load theory, it distinguishes between extraneous, intrinsic, and germane cognitive load. It shows that while AI can usefully reduce unnecessary burdens, it also risks eliminating the productive mental effort – what this paper calls ‘good friction’ – through which counsel and arbitrators build genuine understanding. The analysis demonstrates how friction underpins advocacy, credibility, and strategic foresight for counsel, while ensuring independence and critical reasoning for arbitrators. Overreliance on AI may erode these processes, producing outcomes that are procedurally sound yet substantively fragile. The article also examines the risks of delegating judgment to AI, and argues that without human-framed questions and interpretive frameworks, AI-generated answers remain largely devoid of normative and legal meaning. Ultimately, the article contends that AI should be integrated in a way that removes bad friction while preserving good friction, the effortful processes that transform information into judgment. Only by maintaining this balance can arbitration remain both efficient and legitimate in the age of AI.