Artificial Intelligence (“AI”) is thrust upon us almost every time we are online or use our phones.  It is fast becoming present in many, if not every, aspect of our lives now, and to some degree it is to be welcomed and utilised fully.

There is, however, a major “BUT” to that desire to welcome AI into everything we do, particularly when it comes to the legal world and matters of law. That is the topic we are discussing in this Legal Update, which unfortunately for those involved is based upon very real life events.

Limitations of AI

AI is, without doubt, a brilliant tool and a great time saver if used properly.  However, like any data driven tool, it is only as good as the information upon which it is based and that, in the context of using AI as a means of researching or conducting legal matters, highlights a real need to be cautious as to how much one relies on results that are produced from AI.

The following are examples of where AI has proved to be the undoing of those who relied upon it in the context of legal proceedings and advice.

Where it all goes rather badly wrong

The hapless barrister and their client

The most recent, and highly publicised, consequence of a total reliance upon AI in the legal world arose in relation to the conduct of the barrister in the case of Ayinde -v- London Borough of Haringey, and Al-Haroun -v- Qatar National Bank, in proceedings before the High Court in 2025. 

The barrister in question had, in the usual way, set out in the pleadings for the case the statutory provisions relevant to the case and then cited five separate cases as legal precedent that supported her client’s claim.

The problem, though, was that the quoted statutory provision was misstated and thus misdirected the pleadings and, worst still, the five cases quoted simply did not exist; they were fake!

Whilst denied by the barrister, the court suspected she had used AI to conduct her research and prepare the pleadings. The court said that the barrister had either “deliberately included fake citations” which “would be a clear contempt of court”, or “that she did use generative artificial intelligence tools to produce her list of cases and/or to draft parts of the grounds of claim.” The court went on to conclude that “In that event, her denial is untruthful. Again, that would amount to a contempt. In all the circumstances, we consider that the threshold for initiating contempt proceedings is met”. In other words, one way or another she was taken as misleading and being untruthful before the court. All not very good for the barrister and her client!

The result of the above was that the barrister’s client (a local borough law centre) and the barrister were each ordered to pay a not insubstantial sum under a “wasted costs order” and the barrister faced both contempt of court proceedings and a referral to her regulatory body.  The Judge hearing the case found the barrister and her client to have behaved in a manner that was “improper, unreasonable and negligent”.

All of this because (the court took the view that) the individual concerned had based her pleadings upon information gained via an AI based search of matters relevant to her client’s case.

The lay individual

Whilst we as a firm are all for parties seeking to minimise their exposure to costs by undertaking steps themselves in the conduct of proceedings, we have seen a worrying rise in (our lay opponent’s) use of AI in responding to, or bringing claims against, our clients.

The main concern is the total reliance by the lay opponent on the information gleaned from an AI based research. There is a worrying perception that what is on the internet is absolutely correct all of the time. As most of us know by now, that is simply not the case and so any information found via an AI search must be checked (a point raised over and over by the court in the above case).

We have seen a number of recent cases that we as a firm have been involved in where our client’s opponents use and file with the court responses/defences/claims drafted by AI. Those AI produced pleadings/responses are being relied upon by our opponent as true and correct but often they are wildly off the mark and quote matters of no relevance to the specific matters at hand, or are simply a gross misinterpretation of the law. We have also seen the quoting of non-existent case law precedents, which even the barrister in the case above fell victim to.

In one case the defendant would send to us not only what the AI bot (the one most of you will know) had drafted for him but also the content of what the bot had said to him when providing that draft.  That defendant relied entirely on the AI bot in two separate cases in part because he trusted fully what it said and in part because the “advice” received from the bot was written in a very human and, of most concern, a hugely encouraging manner. Both cases based upon the AI drafted pleadings were summarily dismissed by the court as having no merit or possible prospects of success, in part because the "advice" obtained from the bot was totally and utterly wrong.  

The story is a sad one but far from in isolation and we see an increasing number of people relying upon AI as opposed to seeking proper legal advice from a qualified (and insured) human legal advisers with the expertise to fully investigate a legal matter and apply the relevant legal principles to it. In almost every case, the result is that party incurs far higher costs than might otherwise have been incurred.

Conclusion

At KDL Law we, and lawyers generally, love and adopt new technologies and there is no denying that AI absolutely will, eventually, change the world and how we all live.  BUT, and it is a big "but", beyond perhaps using AI to undertake initial research on any legal matter, you as a landlord, director of a management company, professional property manager or leaseholder/property owner should exercise extreme caution when using AI as a basis for legal advice.

The bottom line is, results obtained from research/advice via AI are not guaranteed to be correct and are not therefore reliable. So that information absolutely must always be checked by a competent qualified and experienced legal professional before that research is relied upon. AI has a value without a doubt, but don’t be of the mind that what you see on the internet/AI bot is absolutely correct as there is a very good chance that it is not and therefore blind reliance upon that could be very costly mistake.

Indeed, it appears that the Tribunals are taking heed of this warning, with the latest ‘Statement of Tribunal Rules and Procedure’ issued by the Southern Region of the FTT now clearly stating that the parties are responsible for checking the accuracy of documents produced using AI and the parties should not assume it is correct. The Statement also requires parties to notify the FTT and the other party if AI has been used to produce or amend any documents. It states expressly that a failure to comply may lead to sanctions including costs orders. This of course demonstrates the caution with which the judiciary are also approaching AI produced information and documents.

The matter is further complicated for lawyers, who have duties of confidentiality and legal privilege attaching to the information they hold or share with their clients, which could become compromised when shared with AI technologies.

No doubt we are still in the very early days with AI especially in the legal world. By its very use, AI learns, develops and evolves, so the future may look very different when it comes to the use of AI tools and technologies. But in the meantime, do not fall into the traps that others have done, but instead approach and use AI in the context of legal advice and proceedings with extreme caution. You have been warned!

Disclaimer

This Legal Update describes the position in law as at the date of this article and care should be taken to note any subsequent amendments to the position as set out above.  The Legal Update is provided free of charge for information purposes only; it does not constitute legal advice and should not be relied on as such. No responsibility for the accuracy and/or correctness of the information and commentary set out in the article, or for any consequences of relying on it, is assumed or accepted by any member of KDL Law or by KDL Law as a whole.