A retiring judge in England tells WIRED that AI could someday decide simpler civil disputes such as personal injury claims

What happened

Geoffrey Vos, a judge in England who was a few days from retirement, told WIRED that artificial intelligence could someday decide some civil court cases on its own. WIRED interviewed him in October at his office in the Royal Courts of Justice in London.

Vos said he recently visited a court in Shenzhen, China, where judges decide civil disputes with heavy help from AI. According to WIRED, he came away impressed and believes courts around the world should consider a similar path. More coverage of these questions is collected in our AI and tech section.

How judges in England use AI now

Guidance limits use to administrative work

WIRED reported that the latest guidance advises judges in England to keep their AI use to administrative tasks. Vos said he uses Microsoft Copilot constantly to find old documents and summarize large amounts of text, but does not use it to write judgments.

He told WIRED he has 11 meetings a day and that Copilot can locate a relevant document created years earlier in seconds. He said he is more cautious about using it to read papers for his cases because he handles some of the most important cases in the system.

A rise in filings tied to chatbots

Vos said the number of county court cases started each year is rising by 10 percent, and that the increase may grow. He linked it to people putting their legal problems into tools such as Gemini, ChatGPT and Claude, which can draft claim documents.

Asked how many of those claims are sound, Vos said the courts do not have the data. He told WIRED it would be a mistake to assume an AI-drafted claim is frivolous, and said the trend shows technology is widening access to justice.

The risks WIRED identified

WIRED noted that several lawyers have been caught citing case law that AI made up. In a first for the UK, someone recently lost their license to practice partly because they misused the technology, the outlet reported. WIRED also said litigants using chatbots as counsel are filing an unprecedented number of claims, some of them bogus.

Deepfakes as evidence

On fake images, video and voice recordings, Vos argued that judging truthfulness has always been part of the job, comparing deepfakes to forged wills. He said people who submit fake images to a court are committing perjury and will go to prison. "I think we need to be cautious about predicting disaster. I'm very optimistic," he told WIRED.

Where Vos would draw the line

Vos said personal injury claims and low-value professional negligence cases could be decided by AI, according to WIRED. He said custody and family cases, along with complex commercial disputes, are areas where people expect human judgment, though AI could help gather information in commercial cases.

"What most people desperately want in civil disputes is resolution," Vos told WIRED. He added that algorithmic decisions would only be acceptable if people consent in a given case or if society decides through legislation.

On the concern that no one can see how a model reaches a decision, Vos said the focus should be on outcomes and on whether the reasoning AI produces is plausible. "I don't think you'll ever do away with human judges," he said, while adding that some people may prefer a fast outcome even if it is not a human one.

What to watch next

• Vos's retirement, which WIRED described as days away, and the handover to what he called a new generation of leadership judges.
• Whether the 10 percent yearly rise in county court cases in England grows further, as Vos said it may.
• Any data on how many AI-drafted claims rest on sound legal grounds, which Vos said courts do not yet have.
• Whether lawmakers take up the question of AI decision-making in courts, which Vos said is for society, politicians and legislation to resolve.