The legal profession is often caricatured as being what you might politely describe as traditionalist in most regards. Admittedly, we do our fair share to cement this stereotype in this blog, and for good reason – spending a career around lawyers, you can’t help but notice a preference for routine and habit, doing things the way they have always been done, and treating change like some kid of infectious disease you want to avoid all contact with.
Which encourages the oft-made assumption that law and technology do not go together, that lawyers will be the last of the Luddites clinging on to the analogue world when everyone else has completed their digital transformation. There is also the argument that a tendency towards technophobia ultimately holds the legal profession back.
But as with most stereotypes, this view is at best an exaggeration. Yes, we’ve met lawyers that won’t even countenance carrying a mobile phone still. But as we pointed out in our last blog, the application of technology in law is a thriving field. So-called legaltech, a whole body of apps, platforms and assorted software designed to aid and facilitate the successful practice of law, is a major growth sector in business technology, attracting huge investment and constituting a sizeable market in its own right.
That being the case, the legal profession can hardly be as averse to technology and change as is sometimes made out. In fact, the notion of using technology to support legal practice is far from new. That legaltech only really took off recently is probably not so much down to lawyers resisting change as the technologies that were available previously not offering enough of a benefit to make their mark.
Law and AI: A Surprising Tale
Artificial Intelligence (AI) makes an interesting case in point here. Widely regarded as being right on the cutting edge of innovation, AI is one of a handful of technologies which many people believe will, after the network revolutions of the internet and mobile, significantly shape human culture for the rest of the 21st Century. Today we have Alexa and those annoying ‘smart assistants’ on our phone that pop up if we press the wrong part of the screen. Tomorrow we may have intelligent machines capable of driving cars, running automated factories and managing city-sized networked infrastructure systems.
A good case can be made for law being one of the first professions to dabble in the use of AI – hardly what you’d expect given its technophobe reputation. As this article points out, legaltech in one form or another has been around since the 1950s. Early efforts were built on the premise of imitating the kind of deductive reasoning and logical decision making – if X applies, then Y follows – that is viewed as the foundation of legal thinking.
So called legal expert systems were seen as a way to speed up and improve the distribution of legal advice. They survived into the 1990s, but the problem was that their rule-based protocols and pre-programmed conditions missed a lot of the nuances involved in human decision-making, such as the ability to handle ambiguity or choosing which rules to apply in which circumstances.
So law got to experience early on the central problem computer scientists have always faced with AI – how to artificially replicate human thought processes. The answer is, don’t – focus instead on what digital computing systems are much better than humans at doing, namely handling and processing large data sets.
Again, legaltech was relatively early in embracing this shift from linear, causative logic to the logic of probability and statistics. Legal ontologies – sometimes given the evocative name ‘legal robots’ – first appeared around 25 years ago. They use data science techniques such as naming, defining and declaring properties of concepts to cross-reference instances across different use cases or within legal frameworks, and attempt to explain their relationships. Their main use is in interpretation.
Although ontologies have remained very much a niche and rather opaque example of technology applied to law, the analytical approach they pioneered can be seen as the precursor to AI-based legaltech as we understand it today. Built on disciplines such as machine learning, predictive analytics and natural language processing, modern AI is far removed from the early reasoning machines churning out their rule-based decisions. It has instead evolved down the path of sophisticated statistical and data modelling techniques, the mathematical science of algorithms, basing ‘intelligence’ on the ability to process large volumes of data to come up with minutely accurate, probability-based interpretations.
AI-based legaltech applications feel like they can think for themselves, just in a different way to your average lawyer.
Lawyers are embracing AI in legaltech for two reasons – it makes their lives easier and, though some would not like to admit it, it can do certain tasks essential to the legal mind much, much better than a person can. For example, AI has been proven to outperform lawyers in checking NDAs for inconsistencies and risks. Not only does it achieve a higher level of accuracy, it can do it much, much faster – in seconds, in fact. Just think of all those hours spent reviewing documents you could save.
This accuracy and speed has all sorts of applications, from legal research to carrying out due diligence. The ability to accurately predict future outcomes based on past patterns in data, and moreover improve accuracy over time by adjusting the modelling through repetition (what Machine Learning boils down to), also means AI can be used to predict legal outcomes. In a dispute, if you know there is an 80% probability that a decision will go against you, it might spur you to accept that first settlement offer.
The worry is, of course, that AI will start to muscle lawyers out of their traditional roles, and indeed Deloitte has made the astonishing prediction that 100,000 legal roles worldwide will be ‘automated’ by 2036. So is the legal profession flirting with its own demise? Unlikely. Law is not all probability and patterns of continuity based on precedent. The very human element of the law is that it changes, it is open to reinterpretation as new contexts arise. AI is therefore very unlikely to replace legal argument for and against, with a judge presiding over the final decision.
Perhaps the reason why AI is thriving and will continue to do so in legaltech is best summed up by this from The FT: “[L]awyers have embraced developing their own technology, such as artificial intelligence, and realised that becoming efficient is not commoditising their services.” Ultimately, lawyers need to earn fees, they need to offer a service clients want and deliver it in a way that is profitable. If AI can help save time and effort to improve the ratio of work to earnings in their favour, most lawyers will surely be won over.