AI will “support but not replace” judicial decision-making, the lead judge for artificial intelligence has said in a newly published position statement.
Sir Colin Birss, president of the business and property division, said the Courts and Tribunal Judiciary’s (CTB) approach towards artificial intelligence will be “cautious but positive” as the technology will be adopted to support judgement but will be subjected to “strict safeguards and robust testing”.
The statement, released last week, sets out how AI will be used, governed and safeguarded within the judicial process, with an emphasis that this aims to compliment the judiciary’s published AI guidance.
Within the statement, Birss spotlighted that AI tools “must not determine the outcome of a case or displace judicial reasoning, analysis or discretion”.
Core principles for the administration of AI are:
- The use of AI must remain consistent with the judiciary’s overarching obligation to protect the integrity of the administration of justice and to uphold the rule of law.
- Judicial decisions will be taken by a judicial office holder. AI may assist, but it must not displace reasoning, evaluation of evidence or the exercise of judicial discretion.
- Judicial independence, impartiality, fairness, accountability, confidentiality and security must be preserved at all times.
- Human oversight is essential. Every AI output used in judicial work must be reviewed, verified and approved by the responsible judicial office holder. The judicial office holder remains fully accountable.
- The judiciary is taking a cautious but positive approach to AI adoption, supported by pilots, evaluation, training, governance and transparency, to maintain public confidence and uphold open justice.
The judiciary has undertaken “substantial work” to develop a “careful and responsible” approach to AI, Birss explained, through its published judicial guidance which has established core principles for responsible use, strengthened guidance on risks including hallucination and bias, and reaffirmed that judicial office holders remain responsible for material produced in their name.
Pilots have examined if AI can support functions including summarisation, transcription, redaction, writing in accessible language, administrative/leadership tasks, chronology-building, and document analysis.
The pilots have also tested reliability of AI tools and their outputs.
Meanwhile only approved AI tools will be used for judicial purposes, the president stressed, adding that their features must include UK data residency, enterprise data protection and disable web access.
Governance and approval over AI will be led by the senior judiciary, who will assign clear responsibility for pilots, approved tools and emerging use cases.
Any new AI tools to be introduced will require “proper approval, evaluation and ongoing review” before wider deployment, Birss noted, and new use cases will be considered by a centralised steering group, chaired by the lead AI judge who will assess legal, ethical, operational, security and data protection issues before proceeding.
This steering group will be the “sole route” for oversight of the development of AI tools as well as a route for escalation should concerns arise
AI training is now provided by the Judicial College, giving guidance and practical support on new and existing tools.
The position statement will be kept under review by the steering group as technology, legal requirements, judicial guidance and operational learning develop, Birss added.
