Polity Atlas
Mechanism · AI & the Future of Governance

AI-assisted arbitration

Dispute resolution in which AI systems triage cases, summarise evidence or propose outcomes, with or without a human decision-maker.

Source attached · not independently reviewedLast verified 2026-10-01 · revision 2 · 1 reviewersRevision historyЧитать на русском

Idea

AI systems help resolve disputes: triaging claims, summarising evidence, suggesting outcomes or, in some proposals, issuing decisions that humans can appeal.

Theory

Proponents expect faster, cheaper justice for small claims. Due-process scholarship insists on a reasoned decision, an accountable decision-maker and a meaningful appeal to a human.

Real examples

  • China established internet courts, starting in Hangzhou in 2017, which use automated tools for online cases.
  • Online dispute resolution on e-commerce platforms handles very large volumes of small claims.
  • Reports in 2019 of an Estonian 'robot judge' were later denied by Estonia's Ministry of Justice.

Critique

Models can reproduce bias in training data, struggle with novel cases and make it hard to see why a decision was reached. The EU AI Act classes AI used to assist judicial decisions as high-risk.

Results

AI is used as an assistant in dispute processes; binding AI-only adjudication with public authority is not documented in the atlas.

Claims

HypothesisCLM-0064 · checked 2026-10-23

Most AI tools deployed in dispute resolution assist with triage, document review or settlement suggestions rather than issuing binding decisions.

Revision history

  1. 2026-10-23rev 2 — Seed import
  2. 2026-10-23rev 1 — Seed import