Use of AI in the legal industry: should we, shouldn’t we?

Answer

This article was first published on 30 May 2023.

The legal industry is typically viewed as being entrenched in traditionalism. However, with the new ‘millennial approach to life’ and consistent technological advancements, there’s been a generational shift in how best to serve clients in a quicker and more efficient manner.

The term 'artificial intelligence' (AI) can be applied to computer systems that are intended to replicate human cognitive functions. In particular, it includes 'machine learning', where algorithms detect patterns in data and apply these new patterns to automate certain tasks. AI can be used for tasks such as legal research, electronic discovery and document management.

The use of AI is becoming more prevalent within law firms and in-house practices, with the likes of Shoosmiths’ Cia and Allen & Overy LLP’s Harvey adding AI tools as another string to the bow of everyday practice within law firms. These programs tend to achieve quicker results than a human could as certain AI programs can complete everyday tasks like drafting contracts and researching answers to complex queries within seconds. With an offering like this, the requirement for fee earners to be trawling through Practical Law for hours could be removed. In turn, the benefit for law firms using AI is that it could increase the amount of free time that legal professionals have. It’d allow them to engage in activity that they don’t always make time to do but should (eg, business development). In the world of the billable hour, it’s beneficial for the client too; less time spent means a cheaper invoice.

But just because AI can be readily applied to legal work, does it mean that it should be?

AI lacks human judgement and sometimes accuracy, so using AI for tasks that have lots of angles and complexities to them may be more trouble than its worth. AI platforms are only as good as the data they’re trained on. It’s therefore wiser for firms and in-house practices to utilise the AI platforms that law firms have trained as opposed to those that have been influenced and programmed by tech giants.

OpenAI’s ChatGPT, for example, can produce answers that, to an untrained eye, can read as plausible but make no practical legal sense. This is because AI doesn’t understand the legal consequences of its outputs – it produces sentences as opposed to legal advice. In addition, the reply is highly dependent on the question being asked, therefore where any question is slightly ambiguous, ChatGPT doesn’t come back to you to clarify but continues to provide an answer based on what it thinks you mean. So, while it seems beneficial to introduce this kind of platform on the face of it, it could be an extremely costly venture if the bot gets it wrong. The benefit of AI producing quicker results is therefore made slightly redundant by the fact that it’ll take the fee earner time to review each outcome for inaccuracies.

Research tasks that are usually reserved for the more junior fee earner are most likely where AI like ChatGPT could thrive as it’d produce much quicker results. Using AI for research tasks would free up the rest of the junior’s time to get involved with more complex tasks, which surely moulds them into better and more well-rounded lawyers for the future. It’d also give them the chance to get involved in more client contact and business development, an area in which they may lack confidence. The relevant fee earner would still have to review the outcome produced to make sure it’s accurate, but it’d speed up the process more than if they had conducted the research themselves to begin with. In reviewing such outcomes for accuracy, the junior would still pick up vital critical thinking and research skills required but from a perspective of confirming what’s written as opposed to the lengthy task of starting from scratch.

To conclude, while some AI programs can be readily applied to legal work, whether they should be is a different story. The installation of AI within law firms/in-house legal teams as an everyday practice should be approached with caution. At this stage, AI platforms that aren’t programmed by law firms should only be reserved for tasks with less intricacies when the fee earner wants to save time. In any event, fee earners should always keep a close eye on what any platform produces before submitting it to the client.

Yasmin Brown is a trainee solicitor at Shoosmiths.