On 6 December 2023 Matthew Hill, Chief Executive of the Legal Services Board, delivered s keynote speech at the Westminster Legal Policy Forum event on artificial intelligence in legal services. His speech is below. Check against delivery.
Speech by Matthew Hill, Chief Executive of the Legal Services Board
Good morning everyone and thank you chair for that kind introduction. It’s a great pleasure to here again participating in yet another timely and informative event organised by the Westminster Legal Policy forum. It’s a particular honour to be sharing a platform with Professor Susskind who is a giant in this field. To be honest I feel like a bit like a teenager guitar enthusiast in the front row of the rock concert being hauled up to play with the Rolling Stones. I hope I can play the right notes, in roughly the right order.
Anyway, to business.
I am going to talk about three things today.
The first is the briefest and I will get it out of the way. AI is here, and has been for some time and in all likelihood – one might say inevitability – it will come to fill every niche of our lives with increasing sophistication and capability. And yet much of the debate around AI has focused on its known current flaws (and often based on the shortcomings of widespread publicly available AI tools of which ChatGPT is arguably the best known).
Such is the pace of development that the public policy debate ought in my view to be centred on the concept of fully capable – perhaps even flawless – AI tools developed for specific purposes, tasks and even sectors. Even if they are not here now, you can bet the house that they will be soon, indeed on timescales very much shorter than the average professional career. In just the four and a bit years I have been in my role the tone of the debate has gone from “AI is all very well but it will never replace lawyers” to “what will happen when we replace all the lawyers”?
That’s the backdrop. My second point and third points will form my main thesis for today. They are, in turn, the regulatory principles that one might apply to the deployment of AI in the law; and the broader potential implications for the way law or legal business is done.
Starting with regulatory principles, at the LSB our overarching approach to technology and innovation is to support its use to deliver public good, improve access to justice and help address unmet legal need, and ensure that regulators and regulation are actively open to innovation to deliver those benefits. This reflect, I think, Richard’s analogy about medical advances and patients. We are looking at AI from the point of view of the consumer and the benefits it might bring in terms of access, cost and convenience.
What do I mean by being “actively” open? Well, we know from our research and engagement that innovators can be deterred from the legal sector because of their perceptions of the complexity of the regulatory landscape. Often, however, innovators are surprised by how receptive legal services regulators are once the ice has been broken, so to speak. So we think that legal services regulators should be actively encouraging innovators into their space, rather than passively waiting for them to make contact, and flexing their regulation to accommodate new approaches, including technological ones, to the delivery of legal services.
To this end, we are developing statutory guidance to regulators on the promotion of technology and innovation that is likely to reflect this “actively” open stance.
Looking at the adoption of technology in the sector, it is fair to suggest that legal services have not exactly been in the vanguard of innovation. In our latest survey we found that for some lawyers and firms even using video conferencing was still relatively new, with most adoption of that now commonplace technology taking place post 2020. Only just over half of firms are using cloud-based storage, and less than a third have a website with interactive features (like a chat function); it seems mildly comical to be talking about basic functionality like this in the context of a debate about AI, but that is the state of play.
When we look at the more advanced end of the scale, only 2-5% of firms are currently using things like blockchain, tech assisted review, or robotic process automation. However – looking forward, up to 20% of firms expect to start using this type of tech in the next three years.
The reason I mention this is that when the lightning pace of development of AI meets the historic innovative inertia of legal services it creates a risk that it is “done to” rather than “led by” the sector.
It is to mitigate this risk that I think it important to build consensus around principles to underpin the deployment of AI in the sector.
Principles that would underpin a framework and sustainable implementation of new technologies as they come along, rather than a rigid and reactive model that might stifle benefits by obsessing disproportionately about risk.
As a starter for ten, we would propose that any set of principles starts with transparency. It ought to be clear to a client whether and how AI has been used in developing the product – usually the advice – they have purchased. This is not solely transparency for its own sake. If a client knows, for example, that AI has been used to deliver a product in ten seconds that might in the past have taken ten hours, there might be an interesting conversation about how much the client wishes to pay. A bit more on that later.
Another obvious principle would be, in my view, accountability. Sometimes things go wrong, however good the tech or indeed the lawyer. The client will always need to know who is responsible when things do go wrong, and how to seek redress. It will not be good enough to blame the machine.
A further, related, principle, might be something I would clumsily label “explainability”. Perhaps one of the most important expectations of a future legal professional is that he or she understands the advice that is being offered, even if it is generated with AI assistance, and can explain it well to the client. Indeed, we know this is an issue of current concern to the legal profession.
In connection with this, perhaps there is a further principle to be elucidated about “using the right tool for the task”. Much of the public debate around AI, including its deployment in a legal context, has been generated by the huge splash that has been made by ChatGPT. We hear exciting and newsworthy examples of case references being fabricated out of nothing by the AI, and stories – which we should be worried about of course – about the inherent bias in systems that are trained on data generated by humans. But we also need to remember that tools like ChatGPT are not designed to give legal advice. Just as I would not expect good results if I cut my toenails with a chainsaw, we probably ought not to be too surprised if ChatGPT throws out some dodgy advice. What we should be much more surprised about is legal professionals relying on ChatGPT in the first place. Use the right tool for the job should be a watchword.
I am sure that there are lots of other suggestions for governing principles, and I hope we get some emerging from the discussion. Indeed, what I have just said needs to be put in the context of the economy-wide principles proposed by Government.
My third and final point is not related directly to AI but is really about what it will mean to be a lawyer, or to run a business in the law, when much of what we have traditionally thought of as lawyering – the research, analysis and overlay of experience and judgement to synthesis argument and advice – will certainly be capable of being done faster, cheaper and better by machines than by people.
What will be left for the human lawyer to do in this world? There is a sort of Elon Musk view of the future in which, unencumbered by the need to work, humans descend into some sort of Wellsian Eloi existence, in which, shorn of purpose and the need to think, we devolve into a diminished version of humanity.
I would like to take a more optimistic view. Currently there isn’t enough lawyering to go around. Too many people and businesses, even in our mature democracy, do not have access to the legal services that would help them get fairer outcomes.
Maybe the great opportunity that AI brings, is to bring the law fully to the people by drastically cutting unit labour costs, increasing convenience, removing bottlenecks etc. Maybe the great lawyers of the middle future are the ones who capitalise on their human communication skills to act as the trusted navigator. The ones who understand both how the technology is best deployed in any given case AND how to articulate the problems that people are living through in a way that gets the best out of the technology. I don’t know.
But coming back from my flight of fancy, I think there is a rather more predictable set of considerations for the nearer future that none of us is really getting to grips with fully. My list would include:
- What are the implications for training, education, onboarding, competence and standards? We currently train lawyers for the jobs of the last fifty years rather than the next 50
- Does automation risk de-skilling? Does that matter? The industrial revolution spelled the end of traditional lace making, for example, but not many of us lament that loss deeply and every day
- What about business models? The writing might be on the wall for the billable hour when work will be done in seconds. Maybe this will mark the transition to a potentially very positive move to value rather than time-based charging (which might in turn have positive effects on diversity in the sector by reducing the reliance on long working hours)
These and many other questions need to be at the heart of a fairly urgent and inclusive debate involving all those with an interest. I very much hope that you will help fill in the many blanks in today’s and future discussions.
I will leave my remarks there.
