• The paper investigates whether AI can legitimately assist or replace human judges, particularly in expressing empathy—a quality the authors argue is essential for procedural justice and legitimacy.
• It defines judicial empathy as a "disciplined capacity to grasp the practical stakes of a dispute" and treats it as a component of a fair hearing and reasoned decision-making.
• The authors propose a governance and design framework for judicial AI based on four principles: human responsibility, procedural transparency, contestability/auditability, and proportxonality of use.
• The research draws on legal doctrine, interdisciplinary scholarship, and insights from a July 2025 stakeholder roundtable.

‍