AI and the future of the legal function

AI and the future of the legal function

AI unquestionably has enormous potential to shape and enhance how General Counsel and the functions they lead operate and
demonstrate value. But there is, and will likely continue to be, a pronounced gap when it comes to the nuanced actions and
decisions which demand high levels of judgement, creativity, and empathy.

As part of our Next Generation GC Network, Hedley May had the pleasure of bringing together senior legal leaders in a
fascinating and incisive discussion about what a technology-enabled legal function of the future could look like. Here, we
summarise some key themes that arose:

We’re at an inflection point – generative AI is fundamentally different

Though the potential for AI to transform the legal profession has been widely acknowledged for years, our participants agreed
that new technological capabilities enabled by ChatGPT and competing large language models (LLMs) represent a fundamental
step change.

As one put it, “We’ve seen legal tech over the years that had very slow adoption – as well as lots of legal tech providers simply
putting an ‘AI sticker’ on top of whatever they’re selling. What we’re talking about now is something completely new. Generative
AI is a different beast”.

This includes, for example, the increasing feasibility of using AI to generate drafts based on prompts, to interpret legal provisions,
to reason about information, and to act as a sounding board for human thinking – capabilities which, until recently, were far
outside the scope of what was considered possible to automate. One of our participants drew parallels between the coming
impact of generative AI on the legal profession and the way it is already transforming the field of computer programming –
whether in code or natural language, technology can increasingly draft to precise standards when given well-defined instructions.

It’s not a question of ‘if’, but ‘how’…

While there was a broad consensus that AI has irrevocably changed the game, how it will be used by legal functions in the near term remains uncertain. A major challenge for leaders is working out their comfort levels around which human activities AI will be used to replace, which it will augment, and to what degree its outputs can be trusted.

The general feeling in our meeting was that at present, AI will largely be used to automate repetitive work, and augment – but not take the lead on – more cognitively-demanding work. One participant said: “It’s potentially very valuable because it takes a different look at data and is able to put it together in a way that may be different from how you would have done it”. Others pointed to weaknesses in AI compared to skilled human lawyers – such as often lacking contextual awareness, and ad hoc unexplainable errors – that point towards technology not yet being ready to take the reins.

The scale of the opportunity means that legal functions, and by extension organisations, that fall behind their peers in AI adoption could find themselves at a serious competitive disadvantage. But adoption requires wide buy-in, among people who are often very accustomed to working in certain ways. As one of our participants put it: “We’re going to have to bring our teams with us, because I think they’re going to be initially quite sceptical. And then we’re also going to have to be the advocate to the business to get them excited and comfortable with using the technology”.

The challenges of training junior lawyers.

If AI is used to automate basic legal work, could it inhibit the training of junior lawyers in rudimentary but vital skills? These skills are often gained by performing relatively routine and mundane tasks, in order to build a baseline understanding of documentation and legal issues – upon which more complicated judgement calls are based.

“How would you get junior lawyers to a place where they’ve got five years of quality and experience, but they haven’t had to do five years of pretty mundane work because technology is now doing it all?”, asked one participant. Another observed: “The easy legal opinions are also the way for more junior lawyers to learn how to do the difficult ones. If you take that away, what’s the impact?”.

However, our group of leaders did not see these challenges as a persuasive reason to avoid the use of AI, when it can markedly improve the efficiency of repetitive work and curtail workloads. Rather, the middle ground involves training junior lawyers to incorporate new technological tools into their workflows while not becoming overly reliant on them. A crucial skill is to be able to identify when and why these tools might make mistakes, rather than just blindly following their outputs.

Legal functions need to prepare for the pace of technological change.

A recurring theme throughout our discussion was about the startling speed at which AI capabilities are advancing – which was a cause for apprehension as well as excitement. “There’s a slight conceit that we’re looking at how we can leverage Chat GPT and AI, when in actual fact I think it’s going to be forced on us”, said one of our participants, comparing the scale of impact to a hypothetical company that has never heard of a computer being suddenly furnished with cutting-edge laptops. “You have to say yes, and so you have no choice but to adopt it.”

One concern is that regulation around AI will fail to keep up with the pace of change – which may force firms to devise their own internal rules, for example, around AI safety. There was also a feeling that AI might over time begin to exceed human performance at more cognitively complex tasks, such as nuanced decision making.

The optimistic view is that further technological advances will leverage General Counsel and legal functions in their focus on strategy and high-value tasks. One of our participants said: “We’ll need to be able to pivot to analyse the data that the AI is spitting out, and understand, interpret, and then communicate these results. That’s when our skillsets and our judgement will come into play even more than ever before”.