ARIHQ v. Santé Québec and the New ICC Rules: What International Arbitration Needs From AI Governance
Earlier this year, the Québec Superior Court set aside an arbitral award in ARIHQ v. Santé Québec after finding that the authorities cited in its supporting submissions had been fabricated by a generative AI tool. It is the kind of case international arbitration practitioners have been quietly dreading since AI tools entered everyday practice — and it landed just months after the 2026 ICC Arbitration Rules took effect on 1 June 2026.
The new ICC Rules do not mention artificial intelligence by name. Commentators argue they govern it anyway, through the general duty in Article 46 to act in the spirit of the Rules and to render an enforceable award — a duty read as extending to how tribunals and counsel use AI tools. Meanwhile, the EU AI Act's Annex III classifies AI used in, or similarly to, judicial and dispute-resolution functions as high-risk, though the Digital Omnibus on AI has deferred those obligations from August 2026 to December 2027. The result, for now, is a patchwork: general principles, deferred regulation, and a growing list of institutional guidelines issued separately by ICCA, UNCITRAL, the LCIA, HKIAC, and CIArb.
Artificial Intelligence and the Future of International Arbitration Practice (Fernando Messias, Kluwer Law International, 2026) is built for exactly this gap. Its central contribution is a working framework for distinguishing permissible AI assistance from impermissible delegation of arbitral authority — anchored in equality of arms, due process, and the reasoned-award standard a case like ARIHQ puts to the test. A dedicated appendix reproduces the ICCA, UNCITRAL, LCIA, HKIAC, and CIArb guidelines in full, so practitioners no longer need to track five separate sources.
In practice, the book supports three recurring tasks: verifying AI-generated legal authorities before they reach a tribunal, drafting procedural orders that set out what AI use is and is not permitted during a proceeding, and training newer arbitrators and associates to recognise algorithmic bias and due-process risk in AI-assisted submissions.
For law libraries and institutional collections, this is a title worth acquiring now rather than later: legal-technology and dispute-resolution holdings are thin on AI-specific arbitration material, and this is currently the only volume that consolidates the institutional guidelines alongside an applied governance framework.
CLNZ Books ships this title worldwide, with secure payment by credit card or PayPal, and supports institutional and library orders through our Request a Quote service.
Q&A
Q: Do the 2026 ICC Arbitration Rules regulate the use of AI in arbitration?
A: Not expressly, but Article 46's duty to act in the spirit of the Rules and secure an enforceable award is being read by commentators as extending to AI use by tribunals and counsel.
Q: What happened in ARIHQ v. Santé Québec?
A: The Québec Superior Court set aside an award after finding the supporting authorities had been fabricated by a generative AI tool, making it a leading cautionary precedent for AI use in arbitration-adjacent proceedings.
Q: Does the EU AI Act apply to international arbitration?
A: Its Annex III classifies AI used in judicial or similar dispute-resolution functions as high-risk; the related obligations have been deferred from August 2026 to December 2027 under the Digital Omnibus on AI.
Q: Which institutions have issued guidance on AI in arbitration?
A: ICCA, UNCITRAL, the LCIA, HKIAC, and CIArb have each issued separate guidelines, all reproduced in full in this book's appendix.
Q: Where can I buy Artificial Intelligence and the Future of International Arbitration Practice?
A: It is available now from CLNZ Books, with worldwide shipping and institutional ordering via our Request a Quote page.