In a few days, judge Geoffrey Vos will retire as Master of the Rolls, the most senior position in England’s civil justice system. As one of his final acts, he’s trying to stoke a debate about whether—and how—judges should be using artificial intelligence.

Under the latest guidance, judges in England are advised to confine their use of AI to administrative tasks, and Vos tells WIRED he uses Microsoft Copilot all the time to locate far-flung documents and summarize large bodies of text. But recently, he visited a court in Shenzhen, China, where civil disputes are already being decided by judges with heavy assistance from AI. Impressed by what he saw, Vos believes that courts across the globe should consider following.

That road is littered with pitfalls. Multiple lawyers have been caught citing case law hallucinated by AI, and in a first for the UK, somebody recently lost their license to practice in part because they misused the technology. Meanwhile, with chatbots as their legal counsel, litigants are flooding the courts with an unprecedented number of claims, some of them bogus.

Yet Vos remains convinced that AI will someday help resolve civil disputes more rapidly and in ways that cost far less. While complicated commercial disputes and sensitive family cases might require a human judge, he believes that personal injury claims, low-value professional negligence cases, and the like could easily be adjudicated by AI.

“What most people desperately want in civil disputes is resolution,” Vos tells WIRED. “The question is, where is the line to be drawn?”Given his impending retirement, the debate for Vos is largely a philosophical one. But for everyone else in the profession—and for coders, cybersecurity experts, math professors, and a whole host of others—the rules are changing fast.

In October, WIRED met with Vos at his office at the Royal Courts of Justice, London. We discussed the ways that AI is spilling into the courts, the practical and ethical issues it raises, and how the justice system might adapt to serve a new generation raised alongside the technology. This conversation has been edited for length and clarity.

WIRED: You’re just a few days away from retirement. How do you feel about leaving the post at this particular juncture, as AI begins to reshape how law is practiced and justice administered?

GEOFFREY VOS: When I started as a lawyer, we had foolscap paper, typists, and Tipp-Ex. We moved rapidly into word processors, faxes, and eventually computers.

Lawyers are conservative with a lowercase c. But that doesn’t mean that they should be backward-looking or retrogressive. It means they should be adopting change cautiously, with their eyes open. The legal system is for the people of this country.

You can’t be effective in providing access to justice if you’re not on the same technological wavelength. But it’s time to leave it to a new generation of leadership judges. How have you built AI into your practice to date? I don’t use it for writing judgments, but I use it for finding things out, navigating large quantities of material, and checking things.

It saves me time. I have 11 meetings a day. I know there’s a document created years ago that’s relevant to my meeting—and Copilot finds it in seconds. I do that every day. In law, isn’t the devil always in the details?

Doesn’t confidence in the legal system depend on the understanding that a judge will review and consider every inch of every court submission? I’m a little more cautious about using it to read papers for my cases, but that’s because I do the most important cases in the system.

People have confidence in our justice system because they believe that our judges have absolute integrity.

In order not to lose that integrity, you have to explain how you’re adopting technology and achieve a broad consensus that it is appropriate. How is AI spilling into the courts right now?

Is it changing how litigants and legal professionals behave? What issues is it throwing up? It’s changing everything remarkably quickly.

We see a 10 percent rise in the number of county court cases started every year—and it looks as if this may increase. That’s because everybody who has a legal problem now puts it into Gemini, ChatGPT, or Claude. Ask it to create a particulars of claim, a case, a claim form—it can. What proportion of these claims are based on sound legal principles, though?

How often are people led by a chatbot to mistakenly believe they have a claim? We don’t have the data.

But I think it’s a mistake to assume that just because a claim is created by AI, it’s totally frivolous. In a way it’s a good problem, because it shows there’s greater access to justice because of technology.

One of the problems with justice in the past is that it’s only been available to rich people. If people have a problem sufficient to bother ChatGPT, they probably genuinely think they have a claim.

When you explore their problem in detail, you might find it is actually a bit different, but it may still be a legal problem and may still be reasonable for them to have wanted to bring it to attention. People are not generally time wasters. They’re not generally trying to make trouble.

What about the opportunity for AI to fabricate evidence on which judges have previously been able to rely—images, video, voice recordings? Where does that leave the courts?

Often, judges have to decide whether a person is telling the truth. That is the most difficult thing to do, because humans are not bad at lying. But judges make these decisions every day. There’s absolutely no difference—in scale or qualitatively—in somebody producing a false image created by AI.

The only reason we’re all getting so excited about deepfakes is because they are so new. But actually, that is the stuff of judging.

But we’re talking about a whole category of evidence that’s now somewhat compromised. Doesn’t that affect a judge’s ability to decide whether somebody is lying?

It was always thus. We always had to decide whether wills were forged, because people used to do that. It’s exactly the same with voice recordings, and video. It’s just an extension of the type of false information that is produced to the court.

People will find out when they produce fake images to the court that they are committing perjury and they will go to prison, just as if they’d forged a will. I think we need to be cautious about predicting disaster. I’m very optimistic.

You’ve previously described an “inevitable” future where simpler civil cases are adjudicated entirely by AI. Why are you confident that’s where we’re headed?

You need to find a way to deliver justice that is within the cultural grain. That doesn’t mean you just conduct all your cases on TikTok. But if you don’t understand that people are asking LLMs for the answer to almost every question under the sun, including legal questions, you’re likely to get it wrong when you try to devise a process fit for the 21st century.

It’s all about consensus. It will only be possible to ask people to accept algorithmic decision-making if they consent to the process in a particular case, or if society decides through legislation.

When people go to court and get an adverse outcome, even if they regard the outcome as idiosyncratic or even wrong, when it’s delivered in a human way, it can be accepted over time. With AI, we have to be careful; it may be less easy for people to accept an adverse machine-driven decision.

I very much want to see a very open, international discussion about how we should resolve disputes.

I suspect part of the reason people object to trial by AI is because we can’t know precisely how a model has arrived at a decision. What’s to be done about this problem of explainability?

I don’t think it’s about reasoning; it’s about outcomes. The AI can produce very plausible reasoning. You can’t see how it produced that reasoning, but you can see the reasoning it produced, and you can judge the plausibility.

I don’t think you’ll ever do away with human judges. There’s something almost biblical about the human judge; it’s in the human psyche to want to be judged by their fellow man. But that does not mean that in some situations people wouldn’t rather have a quick outcome—even if it’s not a human outcome. Where would you draw the line?

I don’t think judges are yet in the position to say. It’s a question for society, politicians, and legislation to resolve. That debate is really important.

I think that custody and family cases are one of the areas where essentially human judgment is expected, even by the younger generation. In complex commercial cases, I think that’s also the case. But I think we will whittle away at how commercial cases are decided by using AI to collate information to make the decision-making process better and easier.

We will allow more people access to justice more cheaply and more quickly using technology. That will be an overall good for the economy and for the people we serve.