Richard Susskind recently contributed a column in The Times Why the lawyers of tomorrow are likely to be AI architects (you can read it if you are a Times subscriber). In the summary on LinkedIn, he argues that professors who are debating how law students should use AI must also ask the deeper question – what are we training young lawyers to become?
“One response to this question follows what I call Model 1, which holds that AI will never be more than an advanced tool for lawyers. They should master legal skills, aided and unaided by AI, and sit papers that assess their competence in both.
A different response emerges for those who believe in Model 2 which holds that AGI (artificial general intelligence) is almost upon us. However, AI systems for non-lawyers will not emerge from nowhere. For now, humans will be needed to design and develop these systems. I argue, most recently in book ‘The Future of Law’, that this will be the job for tomorrow’s lawyers – to build and maintain the AI systems that replace traditional legal service. This will call for fundamentally different training, for example, in design thinking, systems design, legal engineering, data science, and more.”
My response to this question is: both! Legal thinking is good at suggesting such either/or dichotomies and present them as inevitable choices. This is an unhelpful approach to an uncertain future. We do not know how things are going to develop, and at what pace. My experience is that developments usually take longer than we can foresee. Nor will there be a big bang change from Model 1 justice to Model 2 justice. I even predict that Model 2 will still have elements of Model 1 in it.
Clearly, future lawyers will need Model 1 training, provided they will also learn to think for themselves. I recall speaking with a group of them in a country where independent thinking is not always appreciated, to say the least. To get them to speak up and join a discussion in the meeting was extremely hard work.
Meanwhile, training people to design and develop Model 2 systems that may replace, partly or entirely, traditional legal service needs people with different skills. This requires training in different competences, such as data engineering, information science as well as legal competences. That makes it a different ball game from traditional legal training. This kind of training will also attract people with different interests and mental makeup.
I just visited the information organization of the Netherlands judiciary on a social visit, and I already see both models working side by side. The judicial academy trains judges and court staff in writing good prompts. Elsewhere, multidisciplinary teams work on AI-supported systems for non-lawyers. In a way, the future is already here.
I am sure Richard Susskind very much agrees with me.
