'AI unlikely to replace lawyers but instead enhance their service'

Take the internet, email or mobile broadband; all so firmly fixed in our daily existence now that it is hard to imagine our work or personal lives before they came along.

Generative artificial intelligence perhaps offers similar potential. Early versions are showing us a tantalising glimpse of what is to come, with AI-generated content (music, artwork, computer code, prose, commentary) already proliferating.

In knowledge-based professional services such as law, there is a world of opportunity. Bold claims about AI’s ability to dispense legal advice, draft contracts, judge disputes, and even to replace lawyers abound. The hype is well and truly building.

The irony, though, is that bold (and sometimes theoretical) claims often stifle a technology’s adoption in its early stages.

The reality is that generative AI’s application in law is still juvenile.

Over-inflated expectations can scare, confuse and disappoint when the reality does not match up. That is the danger for generative AI and law.

The tech is impressive and there is exciting potential, but it does not necessarily live up to the grand and sweeping statements being made about it – yet.

The reality is that generative AI’s application in law is still juvenile. Its outputs need a lot of checking, for example.

We know about the misfires and hallucinations, but in practice an answer that’s 50 per cent to 60 per cent accurate on average needs a significant amount of work before it can be relied on. A lawyer’s input remains critical.

Currently it is of good use for certain tasks like summarising, simple analysis, drafting and research, but that is a limited sub-set of a lawyer’s work. It is developing fast and new areas are opening up almost by the week, but we are still at the foothills in terms of its breadth of application.

And AI does not exhibit human qualities like empathy. It can not quantify risk, understand wider personal or commercial considerations, bargaining power or vulnerability.

While that might sound like an obvious point, when using AI in practice you realise just how fundamental those things are to a lawyer’s work. 

Taking these things together, it is hard to see how the tech itself could ever replace a lawyer.

Rather than AI replacing lawyers, we are more likely to see it turbo-charging their expertise.

The real, but no less exciting, potential for generative AI in law lies in augmentation; to support lawyers to do elements of their job better and smarter, benefiting users of legal services via improved efficiency and enhanced service.

Across the legal profession more widely, there is scope for AI to improve access to justice and accelerate dispute resolution. It could make a tremendous difference, but the reality is more 'tech-enabled lawyer' than AI substitute.

One thing that is absolutely clear from exploring generative AI’s use in law is that it is here to stay. The rate of development is phenomenal and it will likely be ubiquitous across the profession over the next few years.

New tools are being launched at a rate of knots, the technology is being incorporated into existing systems and some law firms are starting to develop their own applications.

There is potential for it to assist with more complex legal analysis, drafting from precedents, negotiation support, judicial interpretation and prediction, large scale document review and diligence, to name but a few.

What is also clear is that wider investment is needed, not just in terms of the technology but in terms of the expertise needed in legal practice to fully unlock the potential.

Legal technologists, AI specialists and computer scientists, and data and information experts will all be critical to get the most out of these tools, as will skills training and education. Law firms will need to start building out in these areas if they have not already.

So, rather than AI replacing lawyers, we are more likely to see it turbo-charging their expertise. And wider, complementary expertise will be needed within the profession to leverage it fully.

If we view it in these terms, we might avoid a lengthy dip into Gartner’s “trough of disillusionment”, where the mismatch between hype and reality hinders progress, and push forwards with creative, productive exploration.

And while we might not see AI taking a lawyer’s job any time soon, in the longer term we might well see lawyers who use AI replacing those who do not.

Chris Tart-Roberts is head of Macfarlanes’ lawtech group