Over the last few months the number of mentions of using AI in family proceedings has increased dramatically both throughout my online community but also in the wider family court world. This is mainly due to the prevalence and general accessibility to the public of chat GTP over the last year or so.
ChatGTP is an incredible tool. I’ve used it myself for a few things and found it very useful. It’s undoubtedly brilliant at what it does.
So what application might it have in family law?
Well, the main areas where AI could possibly offer some assistance would be with knowing procedure, looking at precedents / caselaw, statutory law, and writing statements.
Using AI to learn about case / court procedures
This would be quite useful, provided it is definitely looking at the correct procedure or procedure rules. The problem is – how would you know? Court can be complex and there’s an awful lot of ‘what ifs’. If your case is pretty standard then I suppose you could certainly find out what hearing might be next, or what might happen at a particular hearing – for example what happens at a First Hearing Dispute Resolution Appointment (FHDRA).
Using AI to find precedents / case law to support your case
Someone looking for precedents / caselaw to support their case may find AI useful to assist, but generally the only people that would need to use them are law professionals. Litigants in Person are not expected to know caselaw, and there’s a good reason why – unless you know what you’re doing you run the risk of doing damage to your own case. Caselaw is to prove a point, or establish that a precedent has been established that is similar to your circumstances, and you would like to rely on that in your own case. For example you might say “In the case of X v Y, X did this and the court agreed that this was a good thing to do”. Good, right? Not necessarily. Caselaw falls in and out of favour as the current thinking adapts and develops over the fullness of time. Let’s say you find a piece of caselaw and, a-ha, the precedent it sets fits perfectly in what you’re trying to say. You make the point in a statement or in court, but the barrister for the other party knows that that particular caselaw is no longer relevant / has been replaced by more up to date thinking / was found to be the wrong decision later in that case, upon appeal. Well, you’ve just given the opposition barrister an open goal.
Another problem is that the judge has the option to ‘distinguish’ cases when caselaw is presented. This basically means that a judge can distinguish the case before them from the one being mentioned in caselaw by saying that in fact, the two cases are simply not alike enough to warrant the precedent being accepted. Not so bad as the open goal in the previous example, but a waste of precious statement space, or speaking time in court.
Using AI school to find statutory law
Unlike caselaw, statutory law is law in the traditional sense – it’s written down in statute (in law) and has to be followed. Because there’s far less room for interpretation with statutory law then it could be that using AI is quite effective for looking this up, and learning what it means.
Using AI to write statements
I’m going to focus substantially on statements in this blog post, as this is where I have encountered this happening more than in any other situation. Writing statements can be very difficult, and especially knowing how to structure them or even where to start. I know of several people who have turned to AI because of this to assist them in this area.
It is widely accepted that a statement should be clear, concise, relevant and personalised. For more information and guidance on writing position statements see my guide here. For more information on writing witness statements I would advise using this guide.
Reasons I have seen people give for using AI on statements:
· Not knowing where to start
· Not knowing what to say
· Not knowing the law
· Not knowing what people normally put
· Wanting to get the best chance of success
· Wanting to make sure words carry weight
· Wanting to appear informed and knowledgeable
· Assume it would improve the statement
· Assuming judge wouldn’t realise
· Assuming judge would be impressed
However I’m here to say that it would not be worthwhile, and in fact is likely to be viewed negatively and can even be damaging to your case.
You may be thinking - well of course he’s saying that. ‘He’s someone who is paid to write statements.’ It’s true, I am. But I’m not saying pay me to write it (though you can if you want to! Click here) I’m saying writing it yourself is going to get you a far better outcome than using AI.
In fact, I will go as far as saying that the order of preference for who writes your statement should be, in order of increasing chances of its effectiveness:-
1) Professionally written
2) Self written
3) AI written
By its very nature chatGTP doesn’t know YOUR case so it will attempt to use generic information to fill in the blanks. The problem with that is that it has a very real risk of adding in information that isn’t relevant in your case. It doesn’t know what your concerns are, and it’s very easy to lose sight of what the statement’s purpose is setting out to achieve.
People I’ve spoken to have explained to me that they have used AI because they wanted to sound like lawyers, or at least highly informed, as they think this will add weight to their words. They love their kids so much that they want to make absolutely sure they give it their best shot in court, and believe this to be the best way to achieve their goals. This is my point. Sounding smart is fine, but it does you no favours if you don’t understand the things that you’re purporting to say, which is what a statement is.
As I mentioned above the court does not expect litigants in person to be fluent in legalese, be able to quote caselaw / precedents, but more importantly it does not expect to receive a statement full of information completely irrelevant to your case – they just want to hear from YOU about YOUR children.
Each case is different, and so personal, and personalised, that it’s just not worth taking the risk.
People seem to think there are magical words and phrases that unlock success (for instance I would include ‘narcissist’ or ‘parental alienation’ which are hugely overused and misused here) and win you cases – and that Al will provide these. It couldn't be further from the truth, and it's incredibly transparent to those of us in the profession. These things stick out a mile away, so I can only imagine how obvious it would appear to a highly learned judge.
How have the courts reacted to the use of AI?
In a case I’m aware of a judge asked a litigant if they had written their statement themselves and they foolishly replied, very pleased with themselves “no Judge, I used AI”. The judge was very angry when they heard this, asking why on earth they would do such a thing. Their anger was not so much for the lack of specific relevance this would mean for the content of their statement as we have highlighted above, but for another reason altogether.
The judge was concerned about the security and confidentiality of the information. All Child Arrangements proceedings are confidential because the court wishes to keep the child’s name and circumstances out of public reach. When using chatGTP it is unclear what happens with information entered, how it is stored, or whether it is stored and who might potentially have access to that information.
What was the outcome of the case? Well, the litigant was given a stern warning that they must never use AI to write a statement again in the future.
In summary
Use AI at your own risk. At best you will find it somewhat useful for procedure and law. At worst it could lose you your case.
Nothing in this blog should be treated as legal advice. If you wish to speak to a reliable and experienced McKenzie Friend about your case then please contact us here.
Comments