Governance of Artificial Intelligence in Institutional Arbitration Towards a Legal Framework for Confidentiality Protection and Data Security

Abstract

This research addresses a legal question of pressing contemporary relevance against the backdrop of
the accelerating integration of generative artificial intelligence tools within international institutional
arbitration proceedings: namely, the challenge of reconciling the efficiency of such technologies with the
preservation of confidentiality as a foundational pillar without which arbitration cannot function. The
study subjects the existing international legal framework—both legislative and institutional—to critical
examination, revealing a structural deficiency manifested in the absence of binding and consistent
provisions governing this integration.
The research contributes two theoretical and practical proposals: first, the concept of ‘Confidentiality by
Design’, drawn by analogy from the principle of ‘Privacy by Design’ enshrined in the EU General Data
Protection Regulation (GDPR); and second, a model contractual framework governing the use of
artificial intelligence tools in arbitration agreements. The study adopts a doctrinal and comparative
methodology, with detailed analysis of the rules of the International Chamber of Commerce (ICC), the
London Court of International Arbitration (LCIA), the Stockholm Chamber of Commerce (SCC), the EU
AI Act, and the SVAMC and AAA-ICDR principles.