This year’s Annual Conference took place online on 19th June, bringing members together to hear from leading voices on the issues shaping expert evidence today.
Opening the day was Sir Geoffrey Vos, Master of the Rolls and Head of Civil Justice for England and Wales. Sir Geoffrey presides over the Civil Division of the Court of Appeal and chairs the Civil Justice Council, the Civil Procedure Rule Committee, Online Procedure Rules Committee, UK Jurisdiction Taskforce and the International Jurisdiction Taskforce.
In his keynote, Sir Geoffrey turned to a subject that is, in his own words, “a very hot topic at the moment”: the use of generative AI in preparing Expert Reports.
Contrasting approaches
Views on AI to prepare Expert Reports vary by jurisdiction, Sir Geoffrey explained. In New South Wales, Australia, Experts have been banned from using generative AI to draft reports without the court’s permission.
England and Wales has taken a different path: the Civil Justice Council issued a consultation in February proposing that Experts disclose what AI use they have made — beyond transcription or administrative tasks — and name the tools used.
A rapidly changing landscape
Sir Geoffrey traced how quickly things have moved since ChatGPT launched in November 2022.
Today, almost every prospective litigant consults AI before seeing a lawyer, and judges often find AI-assisted submissions from litigants in person more concise and focused than unassisted ones. Law firms are increasingly asked to confirm advice that a client’s AI has already given, for a fraction of the previous fee.
Ignoring AI, Sir Geoffrey suggested, is simply not realistic for anyone in the litigation business — and the UK Jurisdiction Task Force’s forthcoming legal statement on AI liability, due in July, is likely to address the risk of professionals being found negligent for failing to use AI appropriately.
Two guiding principles
Discussing guidance on the usage of AI, Sir Geoffrey noted that responses to the risks of AI — hallucination, privilege, bias, privacy — fall into two guiding principles:
- Every professional remains personally responsible for their own work product, however it was produced.
- No professional may breach privilege or confidentiality by feeding private information into a public large language model.
What does appropriate AI use look like?
Sir Geoffrey offered three further points:
- AI is a tool, not an oracle — users must engage with it and interrogate its output rather than treating a single query as final.
- Much AI use is already invisible and uncontroversial, from AI-generated search summaries to translation tools. This is why the Civil Justice Council’s proposed rule excludes administrative uses.
- There is a critical line between using AI for parts of a report and using it to produce the whole report — the latter being unacceptable.
He also warned that reliability does not scale neatly: a one-page summary is likely to be accurate, but the more material a Large Language Module (LLM) digests, the more likely it is to miss the point. AI, he said, “is not magic, and we, the users, should stop treating it as if it were.”
“If an Expert, a lawyer, an accountant, an engineer, a doctor produce inaccurate or unreliable opinions or use hallucinatory references, they are likely to be liable in negligence, whether they’ve used AI or not.”
Sir Geoffrey Vos, MAster of the Rolls
Looking ahead
AI’s role in litigation will keep growing, Sir Geoffrey predicted. It is already informing — and sometimes making — decisions in arbitration.
Judicial decision-making, however, is different: Sir Geoffrey does not believe judges, HMCTS or the Ministry of Justice can simply hand decisions to machines, partly because Article 6 of the European Convention on Human Rights guarantees a hearing before an independent and impartial tribunal.
He concluded that AI will be used across almost every part of the litigation process, and the legal community needs an active debate on where the red lines sit. For Experts, the priority is to follow the evolving rules of their court or tribunal.
Importantly, Sir Geoffrey said that he does not expect AI to reduce opportunities for Expert Witnesses — it has already driven a significant rise in the volume of cases reaching the courts.
Crucially, Experts need to embrace AI — but in doing so, they must take the time to learn how to use AI tools properly and understand the risks to ensure they remain complaint with their duties to the court.
Recordings from the entire Annual Conference are available here.