AI will be part of tribunals' future - but only with better data, oversight body finds
Artificial intelligence is not a separate project. It's a continuation of the courts reform vision, according to the Administrative Justice Council's working group on digitalisation and user experience, chaired by Caroline Sheppard, former Traffic Penalty Tribunal chief adjudicator.
The group's final report sets out 11 recommendations to improve access, usability, process efficiency and the effective use of technology across tribunals.
Where AI helps right now
Rule-based systems and supervised machine learning can support automated case triage and categorisation. They can check eligibility, flag missing evidence and suggest the most appropriate hearing pathway, so cases are routed promptly to the correct teams and workloads drop to a manageable level.
AI can also review long submissions, bundles and witness statements, highlighting key issues, precedents and inconsistencies. That reduces repetitive review work for judges and legal officers while helping ensure critical information isn't missed.
For backlog management and scheduling, algorithms can improve resource allocation and reduce delay. The catch: accuracy hinges on the quality and completeness of historical data.
Data gaps are the bottleneck
The report flags weak data collection in several areas, including remote hearings and the proportion of cases that end through withdrawal, discontinuance or lapse. It notes that HMCTS does not systematically collect information on why appeals are withdrawn or why appellants stop participating.
The tribunal data strategy is described as only partially satisfactory, with uneven delivery and constraints from legacy systems and fragmented data ownership. It's also unclear whether HMCTS collects the full range of data needed to manage backlogs effectively, and many judges and staff still can't access timely, usable management information.
For context on the service landscape, see HM Courts & Tribunals Service (HMCTS).
Decision support, not decision making
AI can assist with routine, low-risk matters by flagging relevant case law, precedents and procedural rules to support decision-making and drive consistency. Ethical oversight is essential, and AI cannot replace judicial discretion or human judgment.
What tribunal leaders and legal teams can do now
- Appoint a dedicated "user champion" to drive continuous improvement and keep changes grounded in real user needs.
- Start systematic collection and analysis of withdrawn and lapsed cases, including reasons and timing; commission research to understand why parties disengage.
- Audit historical data for completeness and consistency before deploying scheduling or backlog models; fix gaps and standardise fields first.
- Pilot rule-based or supervised ML triage with clear eligibility criteria, a human-in-the-loop review, and a documented override process.
- Stand up an ethical oversight process (governance, transparency logs, bias monitoring) for any AI tools used by tribunals or listed parties.
- Give judges and managers fast access to actionable MI dashboards so hearing pathways, listings and directions can be adjusted in real time.
Several recommendations are explicit: appoint an individual user champion, collect data on withdrawn and lapsed cases, and fund research into why parties exit the process. These are low-regret moves that unlock the benefits of triage, document review and scheduling tools described in the report.
As Lord Justice Dingemans, senior president of tribunals and AJC chair, put it: "The recommendations offer a path forward, ensuring that digital reform continues to strengthen accessibility, fairness and public confidence in our tribunals."
For practical applications, workflows and tools in this space, see AI for Legal.
Your membership also unlocks:
