AI in the courts and your data: Ranjit Shergill explains

Senior judges in England and Wales have set out their position on how artificial intelligence should be used in the courts. The direction is clear: AI can support judges, provided it is used carefully, but decisions stay with the judge. On the latest episode of Legally Sound, the Ansons Law podcast, Ranjit Shergill, Associate Director in our Dispute Resolution team, discusses what this means for the courts, for law firms and for the businesses and individuals whose data is involved.

Key takeaways

• The judiciary sees AI as a support tool for judges, not a replacement for judicial decision-making.
• The two main risks for the legal sector are inaccurate output ("hallucinations") and the security of confidential data.
• Using AI to handle administrative work, such as building a chronology from a large volume of documents, frees judges to focus on the legal issues.
• Law firms remain responsible to their regulator for client data, whichever technology provider they use.
• Before sharing personal data with an AI tool, you need a lawful basis, proper contractual terms with the provider and a clear understanding of where the data is stored.

What have the judges said about using AI?

The judiciary first issued guidance on the use of AI by judges in December 2023, and has updated it since. The latest position statement, published in October 2026 by Sir Colin Birss, the lead judge for artificial intelligence, confirms that AI tools will be used to support judges, but stops short of AI-driven decision-making.

The guidance does not change the rules. It sets out good practice. Judges remain personally responsible for everything issued in their name, and the updated guidance highlights that a private AI tool, Microsoft's Copilot Chat, is available to judicial office holders.

Ranjit sees it as recognition of what is already happening.

"I think judges are using AI. It's just setting parameters."

What are the main risks of AI in the legal sector?

There are two. The first is that AI tools can produce answers that look convincing but are wrong, including references to cases that do not exist. The second is what happens to confidential information once it is entered into an AI tool.

"I think there's two major concerns using AI. One is what's known as hallucinations, it's gonna get it wrong. The second is confidentiality and it's the data. Where's that data going? How are you keeping it safe?"

Both risks come back to the same point. AI output has to be checked by someone who understands the case, and the person using the tool has to know where the information they put into it goes.

Why should judges use AI at all?

Large cases can involve tens of thousands of pages of documents and transcripts. A judgment in a case of that size has to set out what happened, when it happened and who said what. Much of that work is about locating and ordering facts rather than deciding legal questions.

If AI can help a judge assemble an accurate chronology from the documents, the judge has more time for the difficult legal issues and the case law. That is where judicial time is most valuable.

There is also a wider benefit. The civil and criminal courts are dealing with significant backlogs, and anything that safely speeds up the work of the courts helps the people waiting for a decision.

"If AI aids and assists the administration of justice, it's only positive. So no, I'd encourage judges to use AI, but obviously get a process of due diligence."

How should law firms protect client data when using AI?

The safest approach is to use a closed, business-grade system rather than a free public tool. A closed system is one where the data entered stays within the firm's own controlled environment and is not used to train the provider's models.

"We'd use a closed system where our data isn't extracted beyond our own systems and database."

Most of these systems are still cloud-based. Cloud storage is now generally more secure than keeping data on servers in the office, but firms need to know where the data is held. At Ansons, our data is held on Microsoft Azure servers in the UK.

Even with those protections, the data still sits somewhere, and the risk does not go away.

"I think the problem lawyers are gonna have is well, that data's gonna be stored somewhere, even if you put all the protections in place."

Who is responsible if client data is lost?

The law firm. Solicitors have a duty under the SRA's rules to keep client affairs confidential, and that duty cannot be handed to a technology provider. If a provider is hacked or mishandles data, the firm may have a claim against the provider, but the regulator will look first to the firm.

"It's not the third party, your regulator will come after you."

What does data protection law require?

Any business that puts personal data into an AI tool, not only law firms, needs to think about the UK GDPR and the Data Protection Act 2018. In summary:

• Have a lawful basis. You need a lawful basis for the processing. Consent is one option, but it is not the only one, and many businesses rely on contract or legitimate interests instead.
• Put the right contract in place. Where the AI provider processes personal data on your behalf, you need a written contract with them that meets the requirements of the UK GDPR.
• Check where the data goes. Transferring personal data outside the UK is restricted. It is only allowed where the destination is covered by UK adequacy regulations or appropriate safeguards are in place, such as an approved data transfer agreement.
• Tell people. Your privacy notice should explain how personal data is used, including where it is processed using AI tools.
• Assess the risk. Where the processing is likely to be high risk, a data protection impact assessment is required before you start.

Free public AI tools rarely meet these requirements for confidential or personal information. Business versions of the same tools often can, but the terms need to be checked.

What should businesses do now?

• Identify which AI tools your staff are already using, including free tools on personal devices.
• Decide which tools are approved for confidential or personal data, and which are not.
• Check the provider's terms on data storage, data location and model training.
• Update your privacy notice and internal policies.
• Make sure anyone relying on AI output checks it before it is used.

"So it's a bit of a minefield at the moment."

Ranjit and Martin De Ridder discuss AI in the courts on Legally Sound, the Ansons Law podcast. Watch the episode.

How we can help

Our Dispute Resolution team advises businesses on commercial disputes, contracts and the legal risks of new technology. Ranjit Shergill and the team can help you put sensible policies in place before a problem arises.

If you need advice on the use of AI in your business or the protection of your data, please fill out the contact form on this page, email info@ansons.law or call 01543 267988 and a member of our Dispute Resolution team will get back to you.

This article is for general information only and does not constitute legal advice. Specific legal issues always require the advice of a qualified legal professional.