AI and the Judicial Role: Sir Geoffrey Vos on the Future of Decision-Making

The Courts and Tribunals Judiciary has published the keynote speech delivered by Sir Geoffrey Vos, Master of the Rolls, at the Legal Geek Conference, in which he considered the growing influence of artificial intelligence (“AI”) on legal services and dispute resolution.

Sir Geoffrey observed both members of the public and legal practitioners commonly turn to AI to research legal questions. The judiciary in England and Wales has access to a version of Microsoft Copilot, while judges in South Korea, Kazakhstan and China regularly use AI to support their determination of cases.

Judicial principle and economic pressure

Sir Geoffrey identified a tension between the established principle that judges should read the case documents before reaching a decision and the practical demands of delivering justice. Economic considerations may favour reliable technology that reduces judicial workloads, helping courts address backlogs and resolve disputes at lower cost.

Developments in China

The Shenzhen Intermediate People’s Court has developed an AI system that identifies the principal evidential points, sets out the questions requiring determination and proposes draft reasons and a possible order for the judge’s consideration. According to a White Paper issued by the court in October 2026, judicial work that previously occupied an entire day can now be completed in “tens of minutes”.

However, Sir Geoffrey highlighted a tension between aspects of the White Paper and a fundamental limitation of AI-driven efficiency: lawyers and judges achieve meaningful time savings only if they rely on the technology’s output, rather than independently undertaking all the reading and drafting they would otherwise have performed.

Could AI take over judicial functions?

In Sir Geoffrey’s assessment, an absolute assertion that AI could never replace judges overlooks several practical considerations:

Public demand for faster justice: Significant pressure may emerge to deploy AI to reduce court backlogs and expedite decisions.
Greater reliance on automated analysis: Judges may consequently spend less time reviewing the underlying papers and depend more heavily on AI-generated factual summaries, issue lists and suggested outcomes.
Acceptance in smaller disputes: Parties to lower-value cases are already open to technology-based resolution.
Economic incentives to settle: Once AI becomes sufficiently dependable, parties, at least in smaller cases, may use its predictions of judicial outcomes to guide settlement and may face financial pressure to resolve disputes on that basis.
An uncertain dividing line: It is difficult to define precisely where administrative or mechanical assistance ends and substantive participation in decision-making begins.

Sir Geoffrey warned that “we may be on the edge of a slippery slope”. Parties to straightforward, low-value disputes may willingly accept AI adjudication because it is quicker and cheaper. The challenge for courts is to determine where the boundary should lie between those disputes and cases of greater significance.

International enforcement and future standards

Sir Geoffrey also stressed the importance of avoiding disagreements between jurisdictions over the cross-border enforceability of AI-assisted decisions, particularly where objections concern whether those decisions were made by an independent and impartial tribunal.

Looking ahead, he indicated that the Standing International Forum of Commercial Courts is likely to develop model rules governing the use of AI in commercial decision-making.