In two recent judgements the use of words in pleadings and witness statements have been considered. They are briefly discussed below.

In Spurgeon & Ors v Capita PLC, Master Dagnall commented that unclear language should be avoided in pleadings. This is a stark reminder of one of the first principles learnt in contract law – clarity is key.  

BACKGROUND

The Defendant suffered a cyber-attack following which, some 3973 individual Claimants brought a class action on the basis of breaches of data protection law.

Proceedings were issued using a single claim form and single particulars of claim (PoC). The PoC presented the claimants’ damages in a tabulated format which comprised of lists of alleged distress suffered, a list of claimants, and an ‘X’ marked under each claimant’s name correlating to the alleged distress suffered. The claimants’ solicitors had used words such as “tormented” and “violation of trust” in the list of alleged distress suffered. These words had not come from the claimants’ themselves but from their solicitors. The words were neither expanded upon nor clarified, rather they were used generically for each Claimant.

The Defendant applied to strike out the case. Their position was that the pleadings were an abuse of process and the claimants’ solicitor had acted beyond authority in presenting the PoC in the tabulated manner.

The abuse of process ground was advanced on the basis that the assertions were improperly advanced in the tabulated format and “the solicitors acting for the Claimants … have irrevocably tainted the Claimants’ evidence by ‘putting words into the Claimants’ mouths’ on the fundamental (and highly subjective) issue of loss and damage”. The acting beyond authority ground was not advanced.

DECISION

Master Dagnall found there had not been an abuse of process and the application to strike out was refused. However, he did note the use of words such as “tormented” required clarification and the PoC and tables were unhelpful and should be changed.

In its current format, Master Dagnall found the choice of wording within the PoC and tables “a situation which is undesirable, as: i) What is being asserted in a pleading should be clear ii) The differences in meaning seem already to have contributed to a substantial dispute”[1]. Further, “in the light of what has happened, causes me to regard the use of merely these words as being such as likely to obstruct the just disposal of the proceedings.”[2]

There is no shortage of cases considering the use of AI in generating pleadings and witness statements. Indeed, there is guidance for judges[3] and solicitors[4] on the use of such tools and the Civil Justice Council are currently holding a consultation on Use of AI for Preparing Court Documents[5].

In Father v. Mother (Fact-Finding: DARVO), HHJ Newport stated that if an individual uses AI to draft a witness statement and then signs the statement of truth, they adopt those words their own.

BACKGROUND

The Claimant father (F) in family proceedings used AI to draft witness statements, of which there were 14 in total.  F’s statements for previous proceedings were drafted by solicitors and the latter by F with the aid of ChatGBT. The 14^th iteration was produced for the fact-finding hearing and is the statement that came under scrutiny by HHJ Newport.

Whilst the judge accepted that there is no prohibition against using AI tools, they noted that each iteration of F’s statement contained more exhibits “that amount to no more than ChatGPT answers to whatever F has asked it”[6] and by this hearing, “AI generated material has become an integral part of the statement rather than an exhibit.”[7]

…

“It is difficult to distinguish between what F says and what an algorithm tells F to say. Ultimately, when a witness signs a statement of truth, the contents of the statement are their evidence, and they must speak to it.”[8]   

The judge then went on to say he found little assistance in F’s written submissions.

CONCLUSION

Whilst HHJ Newport stated that he did not intend this judgment to be to be taken as any form of precedent, it appears to be commonsense that AI generated statements must be considered to be adopted wholly as the words of the signatory to the statement of truth. The alternative is unworkable.

With the often-publicised issues such as AI generating ‘hallucinated cases’ and rewording the user’s input the extent that interpretation of the final statement could be construed differently, there is concern that this may have a negative impact on users, rather than the intended positive impact of assisting in drafting a legal document. Not only in the sense that a witness may be considered less reliable if their oral account differs from their written statement or the written statement contains fictitious case law, but also CPR 31.14 is very clear “Proceedings for contempt of court may be brought against a person who makes or causes to be made a false statement in a document, prepared in anticipation of or during proceedings and verified by a statement of truth, without an honest belief in its truth.”

[1] Spurgeon & Ors v Capita PLC [2026] EWHC 241 (KB), para203

[2] Ibid, para 207

[3] https://www.judiciary.uk/guidance-and-resources/artificial-intelligence-ai-judicial-guidance-october-2025/

[4] https://www.lawsociety.org.uk/topics/ai-and-lawtech/generative-ai-the-essentials

[5] https://www.judiciary.uk/related-offices-and-bodies/advisory-bodies/cjc/current-work/use-of-ai-in-preparing-court-documents/

[6] Father v. Mother (Fact-Finding: DARVO) [2025] EWFC 284 (B), Para 101

[7] Ibid, para 102

[8] Ibid, Para 103