When Judges Use AI

Picture the scene. It is 2019, and a judge is dealing with a complicated employment discrimination case.

In front of the judge is a substantial bundle of documents: the original claims, witness statements, supporting evidence and legal submissions from both sides. Before reaching a decision, the judge and their clerks would need to work through these materials carefully. They would research the relevant law, searching legal databases for previous cases involving similar circumstances. They would then compare the legal principles established in those cases with the facts before them, before reviewing the evidence and organising the most important arguments and issues.

This process could involve hours of searching through lengthy documents, identifying relevant passages and separating important information from material that was less significant to the final decision.

Now imagine the same case in 2026. The role of the judge has not been replaced, and the responsibility for reaching the final decision still rests with a human being. However, the process leading up to that decision can look very different. An AI tool can assist with the initial stages of the work by summarising large bundles of documents, identifying relevant passages, organising information and assisting with legal research. Tasks that may have previously taken hours on end now can be done with a few prompts and a click of a button.

This is particularly significant when considering how much of a lawyer's working life has traditionally been spent away from the courtroom. A fascinating 2011 report by the Bar Standards Board, Barristers' Working Lives (England and Wales), found that self-employed barristers worked an average of 53 hours per week.1 Often when one thinks of a barrister, the courtroom advocacy goes straight into mind. However, this study demonstrates that a significant proportion of their work took place at a desk, reviewing extensive document bundles, examining evidence, conducting legal research and drafting court documents.

The change is no longer simply theoretical. A 2026 study conducted by Northwestern University involving 112 judges found that 60% of U.S. federal judges reported using at least one AI tool in their judicial work.2 This suggests that AI is not a technology that courts are merely considering for the future. Judges are already beginning to use these tools in practice.

The significance of this development is not that AI can replace the role of a judge. Rather, it has the potential to change how judges and other legal professionals spend their time. By assisting with document review, summarisation and legal research, AI may reduce the amount of time spent on the initial process of sorting through information. This could enable legal professionals to focus more of their attention on the areas where human judgement is essential such as evaluating evidence, applying the law and ultimately making decisions.

Now that we have established the reason why Judges are using AI and that Judges indeed are using AI, it's crucial to understanding exactly how Judges are using AI.

A study in 2025 was conducted by the National Center for State Courts. The NCSC conducted thirteen one-hour interviews with judges serving across ten different U.S. states.3

The NCSC reported three fascinating takeaways.

A. AI can most definitely support judges, but it cannot replace their decisions. Whilst the judges interviewed used AI in different ways, there was one underlying point they were all clear about: the final decision in a legal case must remain with the judge. AI can assist with certain parts of the judicial process, but deciding the outcome requires human judgment.

B. Judges are using AI mainly to streamline certain tasks. The main use of AI identified in the study was to make certain tasks more effective. As spoken about above, judges and clerks often would spend hours on end organising and filtering through documents, AI has found to potentially help judges work more effectively without changing their fundamental role.

C. Judges highlighted the importance of understanding the risks of AI. The judges placed a strong emphasis on the importance of understanding the danger of using AI in the courtroom. Whilst they recognised the benefits it could offer, they understood these tools are not perfect and they require responsibility.

There is a clear distinction between using AI to assist with judicial work and allowing AI to make judicial decisions.

Across the world, courts are struggling to deal with the large number of cases. This has created delays that can leave people waiting months or even years for their cases to be resolved.

In the UK, by the end of 2025, the Crown Court in England and Wales had more than 80,000 cases.4 The United States also face an enormous caseload. In 2025, the U.S. federal district courts had more than 432,000 pending civil cases,5 while the federal courts of appeal had almost 34,000 cases awaiting resolution.6

These figures represent the scale of challenge facing the court systems around the world. AI cannot solve these backlogs on its own, but by helping review and summarise large amounts of information it could reduce the most time-consuming parts of the legal process.

In summary, the evidence is showing that AI is changing the way legal work is carried out by Barristers and Legal clerks. By helping research and organise large document bundles it is saving barristers a significant amount of time. However, human judgment remains at the heart of the legal process, with limits on how AI is used. Whether these lines are becoming blurred is another question for another article.