How should family lawyers respond when clients use AI?

There was a time when a client might arrive at an appointment with a printout from Google, advice from a friend or something they had read on an online forum. Increasingly, clients are arriving having already asked an AI chatbot about their situation, whether that’s what they could be entitled to financially following separation, or what could happen in relation to their children.

Some clients are using it to help them respond to their former partner, or to work out practical steps to take when separating (such as what should be done with passports). More and more, clients are using AI to help prepare documents for court, turn a long history of events into a chronology, or summarise correspondence they’ve received. Some arrive with something that looks remarkably polished and ask, “Can you just check over this?”

Recent research from JUSTICE and the Administrative Fairness Lab bears this out. It found that around one in six people surveyed who had experienced a legal problem in the previous two years had used an AI chatbot for help, information or advice, and family law was among the areas where people were turning to these tools.

So how should family lawyers respond?

Start with the client’s question

The answer isn’t simply to tell clients not to use any form of AI. It’s readily available and, for someone worried about their children and their financial future, asking a chatbot a question feels like an easy place to start. Previously, they would Google these questions to better understand their position.

When a client does bring an AI-generated answer to the meeting, the instinct shouldn’t be to dismiss it, but to understand how they got there. This means asking what they actually asked the AI tool and what information they gave it. Does the AI know the client is in England and Wales? Has it made assumptions about their circumstances? Most importantly, has the answer changed what the client thinks they should do next? Sometimes the question itself tells us more than the answer, and can help the lawyer understand the client’s concerns.

A question about whether they’re “entitled to half of everything” can be a sign that someone is worried about whether they will have somewhere to live. A parent asking repeatedly about the likelihood of a particular child arrangements order could well be looking for certainty in a situation where we simply can’t give it.

Check what clients bring in

There will be occasions when the information a client brings us is incomplete, misleading or simply wrong. Rather than just telling them not to rely on it, it’s more useful to explain why. Important facts can be missing. The answer might be based on another jurisdiction, present one possible outcome as inevitable or fail to account for the wide discretion available to the Family Court.

The way we check it will depend on what the client has used AI to do. If it’s helped create a chronology, does it accurately reflect the dates and events in the underlying documents, or has something important been missed? If it’s been used to help draft a statement, is the content accurate and does it adhere to the Family Procedure Rules? If it’s summarised a letter or other correspondence, has it preserved the meaning or lost some of the nuance along the way?

These aren’t necessarily new questions for lawyers. AI simply gives us another reason to be particularly careful about where information has come from and how it has been produced.

The same applies to AI-generated correspondence. Something can be beautifully written, but completely wrong for the situation. In family law, a polished letter that unnecessarily escalates conflict can do considerably more harm than good.

Enquire what the AI does with the data already provided by the client

Many clients won’t realise that free AI services may use their data for training and retain the information. Free AI tools may not have appropriate guardrails to secure their personal information. The retention of personal data about their children and family is concerning, especially where the information relates to court proceedings and deeply personal issues. I would suggest asking the client if they are aware whether the AI they are using is secure, and whether the data is retained for training purposes. Clients may not even have realised there is a difference. They would not leave their personal documents on the train for anyone to look at, and the same confidentiality principles should apply with AI.

Ask what clients need from us

The JUSTICE research also found that people weren’t only asking AI for legal answers. In one example, someone had copied an email from their own lawyer into a chatbot and asked for it to be explained in “normal person language”. If a client needs an AI chatbot to understand an email from their lawyer, the question isn’t only whether the chatbot has explained it correctly. It’s whether the email needs redrafting. It is a reminder to lawyers that legal jargon should be avoided.

Family law clients are often dealing with an extraordinary amount at once, worrying about their children, their home or their financial future while also being expected to absorb unfamiliar terminology and make significant decisions. Clear communication has always been part of good practice. AI hasn’t changed that.

Some people are also using AI to prepare for legal advice. In one example from the research, someone asked what to take to a first meeting with a lawyer and what questions to ask.

Clients are going to use AI, just as they use Google, online forums and advice from friends and family. What matters is what happens when they bring that information to us. We can help them understand what is useful, what needs to be treated with caution and where the individual circumstances of their case make a generic answer unreliable.

If we’re prepared to talk about it openly, the questions clients are asking AI might tell us quite a lot about what they understand, what they’re worried about and, sometimes, where we could be communicating better too.

Katie Phillips is Head of Family Law at LEAP Legal Software.

Photo by Vitaly Gariev on Unsplash.