Use it, verify it, own it: AI and the new TCC Guide

On 1 July 2026, the Fourth Edition of the Technology and Construction Court Guide (the “Guide”) came into force – the first new edition since October 2022.  Much of it will only interest litigators.  One section should interest anyone who runs, builds or invests in a technology business: section 1.6, ‘The Use of Artificial Intelligence’.

It runs to five short paragraphs.  They are the clearest statement yet, in an English court’s own procedural guide, of how judges expect AI to be used in litigation – and they double as a serviceable minimum AI policy for any business.

The court where technology disputes end up

The Technology and Construction Court (the “TCC”) is a specialist court within the Business and Property Courts.  The ‘Technology’ in the name is not decorative: the Guide’s own list of TCC claims includes ‘claims relating to the design, supply and installation of computer systems, computer software and related network systems’ (paragraph 1.3.1(g)).  If a major software implementation fails, or a software development contract collapses, there is a fair chance the dispute is tried by a TCC judge.

And the Guide is not commentary.  Parties, solicitors and counsel are expected to follow ‘both the letter and spirit’ of it, and the court may impose sanctions, including costs orders, on those who act unreasonably or fail to comply (paragraph 1.1.2).  What the Guide says about AI is procedure.

Four rules in five paragraphs

Section 1.6 does four things:

• Permits AI use – anchored in personal responsibility (1.6.1). Legal representatives are personally responsible for the material they put before the court and have a professional obligation to ensure that it is accurate and appropriate.  Provided they are mindful of that responsibility, AI can be used.  Note the starting point: permission, not prohibition.
• Defaults to disclosure (1.6.2). Whether, and to what extent, AI use should be disclosed depends on the context – but any doubt ‘must be resolved in favour of disclosing such use’.
• Protects the underlying data (1.6.3). Representatives must at all times uphold the confidentiality, and protect the privacy, of any underlying data in relation to which AI is used.  In practice: client documents and dispute materials do not go into public chatbots, and the terms on which any AI tool stores, processes and trains on data need to be understood before the tool is used.
• Requires independent verification (1.6.4). The accuracy of any research or case citations generated with the assistance of AI must be independently verified – and the representative must ‘be in a position to confirm’ that they have done so.  Expect judges to ask.

Paragraph 1.6.5 then signposts the judiciary’s own AI guidance for judicial office holders of 31 October 2025.  That guidance deals, among other things, with hallucinated authorities, ‘white text’ (prompts hidden in documents so as to be visible to a machine but not to a human reader) and the rule of thumb that anything typed into a public chatbot should be treated as published to the world.

How five paragraphs earned their place

The backdrop is not hypothetical.  In Ayinde v London Borough of Haringey; Al-Haroun v Qatar National Bank QPSC [2025] EWHC 1383 (Admin), the Divisional Court dealt with lawyers who had put fabricated or unverifiable authorities before the courts, with AI tools implicated in how the citations came to exist.  Dame Victoria Sharp, President of the King’s Bench Division, warned that promulgating guidance on its own was insufficient to address the misuse of AI, and that more needed to be done to ensure lawyers complied with their duties to the court.

The system responded quickly.  The Chancery Guide was amended in September 2025 to require parties to ensure that authorities obtained online, particularly via AI tools, are genuine.  The judicial AI guidance was refreshed in October 2025.  In February 2026, the Civil Justice Council (the “CJC”) published an interim report and consultation on whether new rules are needed to govern AI in the preparation of court documents.  The consultation closed on 14 April 2026, and the CJC’s June 2026 update on the responses reports broad consensus that no new AI-specific rules are needed for professional drafting – existing professional responsibility does the work – with witness statements the live question, ‘proportionate transparency’ the direction of travel for expert evidence, and a final report expected later in 2026.

Section 1.6 distils that direction of travel into four operating rules and puts them in the court’s rulebook.

The TCC has form

None of this is a court discovering technology.  The TCC was the first court to develop an e-disclosure protocol, and the English courts’ embrace of technology assisted review dates back a decade, to Pyrrho Investments Ltd v MWB Property Ltd [2016] EWHC 256 (Ch).  The new Guide continues the theme: parties are expressly encouraged to reduce the burden and cost of disclosure through software and analytical tools, including technology assisted review (paragraph 12.1.10).  The judicial posture is not scepticism about AI.  It is AI on documented, verified terms.

And now the squeeze from the other side

The Guide landed in the same fortnight as the UK Jurisdiction Taskforce’s Legal Statement on Liability for AI Harms (7 July 2026), which we at HLaw covered.  The Statement concludes that a professional may be negligent for using AI carelessly – and, in the right circumstances, for failing to use it where a competent member of the profession would have done so.  One of the Statement’s own examples is a solicitor in the Business and Property Courts who fails to advise the client to consider AI-assisted document review.

Put the two documents side by side and the position for litigation professionals is symmetrical.  Use AI badly and you answer for it.  Refuse to engage with it at all and, before long, you may answer for that too.  The corridor left open between the two is exactly the one section 1.6 describes: use it, verify it, own it.

What to do with this if you are (or may end up) litigating

• Ask your lawyers the section 1.6 questions. Which AI tools will touch your dispute materials, and on what terms (1.6.3)?  What is the verification process for AI-assisted research (1.6.4)?  Who takes responsibility for what goes before the court (1.6.1)?  Good answers exist; you are entitled to hear them.
• Keep AI away from witness evidence. Trial witness statements in the TCC must comply with Practice Direction 57AC: so far as practicable in the witness’s own words (Guide paragraph 13.1.6), verified by a statement of truth and certified by the legal representative (13.1.8–13.1.9), with sanctions that include the statement being excluded from evidence (13.1.10).  The CJC has consulted on a declaration that AI has not been used to generate the content of trial witness statements.  A model that ‘tidies up’ a witness’s recollection is a threat to the authenticity that makes the evidence worth anything.
• Ask your experts now. The Guide does not expressly address experts’ use of AI, but most TCC actions rely in detail on expert evidence (paragraph 14.1.2), and the CJC’s direction of travel is disclosure of whether and how AI has been used in expert reports.  Choose experts who can answer that question cleanly.
• Borrow section 1.6 for the boardroom. Responsibility, disclosure, confidentiality, verification.  Four lines that govern the lawyers  appearing before the High Court will serve most businesses as a minimum AI policy – and they slot in alongside the governance frameworks we have written about in the context of the UK Cyber Governance Code of Practice and the Data (Use and Access) Act 2025 .

From ethics column to procedure

Nothing in section 1.6 is conceptually new.  Lawyers were always responsible for what they put before a court.  What is new is the location: these expectations now sit in the procedural guide of a specialist High Court list, published by the Judicial Office, to be followed in letter and spirit, with costs sanctions behind them.  Responsible AI use in litigation has stopped being a topic for ethics columns and become a procedural fact.  In our experience, where the courts land first, regulators and counterparties tend to follow.

If you would like to talk through AI use in a live or threatened dispute, or want your AI governance reviewed against where the courts and the UKJT have landed, contact a member of the HLaw team.

All the thoughts and commentary that HLaw publishes on this website, including those set out above, are subject to the terms and conditions of use of this website.  None of the above constitutes legal advice and is not to be relied upon.  Much of the above will no doubt fall out of date and conflict with future law and practice one day.  None of the above should be relied upon. Always seek your own independent professional advice.

Humphreys Law