Artificial Intelligence – Evolve or Die?

This article gives an account of  increasingly significant impact that Artificial Intelligence has on legal services provided in the UK, especially in the context of SaaS platform and data protection laws.

Artificial Intelligence has to date perhaps been better at generating publicity than effecting major changes in the way that law is practised in the UK. While law’s reputation as an industry of Luddites is no longer deserved, if it ever was, it is fair to say that the use of technology in the delivery of legal services has evolved at a comparatively slower pace than some related professions, such as finance. However, despite the industry’s conservatism, the trend towards greater use of technology is clear. This development is liable to have been accelerated by the recent pandemic, which has familiarised many lawyers with a range of useful new technologies and placed a large number of clients under financial pressure at a time when they may also have an increased need for legal advice. That is liable to mean that law firms find themselves under increased pressure to do more for less, which is likely to be a conducive environment for AI to become a more deeply integrated feature of legal services. This may present significant opportunities, but also novel types of risk, for law firms in the UK.


To date, the applications of AI in legal practice in the UK (and elsewhere) have related to five main areas: (i) disclosure/discovery (i.e. technology assisted review of litigation disclosure), (ii) contract review (e.g. flagging discrepancies in/features of documents, often in a transactional due diligence context), (iii) document automation (e.g. creating tailored legal templates), (iv) legal research (e.g. improved natural language search) and (v) practice management (e.g. automated billing). All of these areas have seen significant advances in
recent years (indeed, many lawyers now complain that there is too much tech from which to choose). However, these advances may be generally characterised as optimisations of traditional processes within law firms, rather than technology changing the way in which legal services are delivered. While assuming a slow pace of change is usually a good bet in law, a number of trends (both digital and otherwise) have potential to open the door to AI having an increasingly significant impact on legal practice in the UK going forwards.

From a technical perspective, AI products (including tried-and-tested enterprise software with AI enhancements) are increasingly cloud-based and accessible. The interest in this area among law firms may be evidenced by the considerable investment that SaaS platform Reyen Court secured from many of the industry’s most famous names before its launch earlier this year, enticed by the prospect of pre-vetted, plug-and-play software on demand. While this may initially be the preserve of international mega-firms, it is possible that smaller organisations with tighter margins are those that stand to benefit and change most from the potential for greater cost and time efficiency that legal SaaS platforms offer. This may be particularly the case if costs of AI applications reduce, as may plausibly result from increased marketplace competition between lawtech developers and, perhaps most significantly, state-of-the-art foundational technologies such as Google’s natural language processing BERT model being available on an open source basis.

From the perspective of broader trends, the realisation that technology has made offices dispensable has been one of the big business stories of the past months. Health concerns and home comforts are liable to make many lawyers take a closer look at what AI and software more generally can do to make their professional lives easier. Greater familiarity with technology may be the most significant driver of change including, perversely, by duty of a better understanding of the limitations of AI, which many professionals may find easier to accept as, e.g., a nifty way of avoiding timesheets than as something more profoundly disruptive. Many law firm clients are now under considerable financial pressure and correspondingly likely to be intolerant of the inefficient use of fee-earner time spent doing tasks that could be done (often better) by machines. In addition, the legal profession in the UK was liberalised towards the end of last year, with various novel forms of (in some cases
unregulated) legal practice being approved, with a view to increasing the number of freelance lawyers offering affordable legal services to the public. With AI-driven systems being at least to some extent capable of replicating traditional infrastructure benefits of a physical law firm (e.g. by handling tasks such as bookkeeping, bundling and bills), it is plausible that an increasing number of practitioners will choose to set up on their own and, e.g., seek to market themselves on price and rely on AI to do many essential non-client facing tasks.

However, the use of AI in law is not without potentially significant issues. Within firms, concerns about the privacy of sensitive client data has been a traditional impediment to technological update and remain a significant concern. The Data Protection Act 2018, the domestic legislation implementing GDPR, introduces both serious civil and criminal liability for types of data breach (although it is possible that the UK’s departure from the EU will lead to a relaxation of laws in this area, consistent with the government’s desire to maintain
the UK’s position as a global leader in AI). In addition, and with relevance to both client and law firm operations, certain types of complex AI, such as machine learning, may create tricky legal issues, e.g., due to the difficulty of establishing causation in circumstances where reverse engineering AI processes to determine why things would have gone wrong is impossible and/or the difficulty in apportioning liability for damages in circumstances where the AI’s output may be a function of the complex interaction of code and data, and the responsibility of multiple potential parties. It may be that these types of problems, together with the traditional conservatism of the legal industry, ensure that AI remains more hype than reality throughout the 2020s. However, there are perhaps more reasons for thinking that its time has come.


Authors