The Courtroom Algorithm: Why AI Cannot Replace Judges, Arbitrators and Other ADR Practitioners

Artificial intelligence is bringing many benefits to the legal profession, streamlining processes, analysing case law, and is even being used in certain jurisdictions to predict outcomes. But as someone who has spent years in dispute resolution, particularly in arbitration and mediation, I think our profession is at considerable risk if it continues to underestimate the potential negative impact of AI if it’s allowed to play too great a role in judicial decision-making.

AI, by its nature, is a prediction tool. It analyses historical data and identifies patterns, but it doesn’t reason or question. It doesn’t pause to consider whether an outcome is just. If previous judgments or arbitral awards were flawed – perhaps influenced by unconscious bias – an AI model will simply carry on repeating those patterns. It won’t correct them; it will reinforce them.  

We must remember that the legal system in the United Kingdom is not just about applying precedent. It requires human judgment, an ability to probe, to challenge, to read between the lines, which is different from how things are done in civil law jurisdictions. A judge observing a case unfold in real time may notice subtleties that an algorithm never could. A witness who repeatedly requests adjournments might be stalling for strategic advantage – perhaps to push a company into insolvency before a ruling is enforced. A human judge or arbitrator would spot this. An AI, lacking context, would not.

There is also a deeper concern: if AI gains too much influence in judicial decisions, will we still recognise justice when we see it? The role of the judiciary is not just to process cases efficiently but to weigh moral consequences, to ensure justice fairness, and to uphold the rule of law in a way that no machine can replicate.

Of course, AI has its place in legal practice. It can help with research, improve efficiency, and support decision-making. But we must not allow it to replace the human mind in matters of justice. There is a remarkable difference between how the human brain and a machine work.

The law is not just about logic – it is about fairness, equity, and humanity. But ultimately, it is about justice. Lord Howard put it best as far back as 1924 in R v. Sussex Justices ex-parte McCarthy, when he said: “Justice should not only be done but justice should manifestly and undoubtedly be seen to be done.” That is something no algorithm can ever fully grasp.