ABSTRACT Mediator discretion in community mediation is inevitable yet often invisible, which can obscure legitimacy, accountability, and bias in publicly funded or state‐adjacent settings. This conceptual design paper uses integrative synthesis, drawing on street‐level bureaucracy, procedural justice, and dispute system design, to derive a practice‐ready and auditable way to govern within‐session discretionary process moves without disrupting conversational flow or capturing confidential narratives. I define transparent mediator discretion as the practice of making discretionary process moves visible through brief reason‐giving, authorizing them through bounded party choice, making them reversible through a normalized reset pathway, and documenting them through minimal coded entries. The paper's results are design outputs: a fairness‐by‐design framework centered on a brief rationale‐and‐consent loop, a proportionality ladder, a menu‐based set of move options, and a code‐only documentation structure. Worked examples illustrate how the loop can be delivered as a human micro‐competency and supported by digital decision aids without replacing human judgment. I also specify falsifiable propositions, privacy‐minimizing indicators designed to avoid narrative capture, and low‐cost evaluation designs that can be implemented in low‐resource programs. Next steps include piloting in both voluntary and court‐connected contexts, testing safeguards for stalemate and power imbalance, and assessing how AI‐assisted support can preserve party choice, reversibility, and accountability under human oversight.
Martin Magmarigen Kwan Ken Wong (Mon,) studied this question.