Darrow AI’s Post

How should we design AI arbitrators? This is the question I had the pleasure of discussing at ODR Cyberweek 2025. While exploring “How might different training methods affect fairness in AI arbitration?”, my key takeaway was the importance of connecting technological design with legal reasoning. I explored how core legal principles can (and sometimes cannot) be built into the way machine learning systems are designed and make decisions. During the discussion, participants raised insightful questions about: (1) The value of combining training methods and; (2) The potential of AI agents to reduce systematic bias. 💡 I’d love to hear your thoughts on these as well! ➡️ Real-world example: link to American Arbitration Association’s first AI arbitrator in the comments. ____ A special thank you to the National Center for Technology and Dispute Resolution (NCTDR) for organizing such an inspiring event, to Colin Rule for the warm invitation and opportunity to share my work, and to Prof. Niva Elkin-Koren, who introduced me to this fascinating field and guided me throughout the process.