This session examines the accountability questions raised by the growing use of AI tools in arbitral proceedings, by counsel, arbitrators and institutions alike. Under the EU AI Act, AI systems used to assist an arbitral tribunal in interpreting the facts and the law are classified as high-risk (Annex III); counsel’s and institutions’ own use of AI generally is not. The timeline for those high-risk obligations is a live issue: originally due to apply from 2 August 2026, they have been deferred to 2 December 2027 under the ‘Digital Omnibus’ package adopted in mid-2026, even as the Act’s AI-transparency duties still take effect on 2 August 2026. The panel will consider who bears responsibility when an AI-generated work product contains errors, what disclosure obligations counsel and arbitrators should have when relying on AI, whether arbitrators may delegate any part of their reasoning process to AI tools and how practitioners should navigate the new regulatory framework in cross-border proceedings.