How will software transform the legal industry?
A general counsel, Jeff Carr, once famously described the things that lawyers do as falling into four buckets: content, process, advocacy and counselling. He then went on to upset all private practice lawyers by saying that he wasn’t going to keep paying for either content or process.
Until Google invents either general artificial intelligence (in which case it’s game over) or Robocop (which will make for better TV), automation of the legal industry won’t be a dramatic change, instead it will look like a rising tide of software that keeps eating more and more of these four areas.
The law is notorious for its lackadaisical approach to replacing wigs let alone technology, so you might – and most lawyers certainly do – think that automation is rare and nonthreatening.
Meanwhile in 2013 start-ups in the legal tech industry raised more than $400m. Big changes are coming fast, just not from the usual quarters.
Software is eating content in three ways: categorising it, evaluating it and tailoring it.
With categorisation, the rise of machine learning has been most visible with predictive coding in litigation. Effectively, software is now demonstrably as good as, if not better than, humans at reviewing huge piles of emails to determine which should be disclosed as relevant to a case. Applying similar technology to the world of contracting and compliance, software can now reach into documents and pull out the relevant provisions or tell you how far your contract deviates from market standards.
Finally, software can now automate the creation of tailored contracts and other content. Once the preserve of enterprise systems requiring programming to creating the specific content, editing has become easy enough to be used by reasonably savvy end users.
And this is only spreading: a recent iOS application allows end users to generate tailored contracts and, having raised significant funding, is going to keep adding contract management features such as allowing the user to manage their agreements from their phone.
Meanwhile when it comes to the content itself, new providers are selling legal content at commodity (i.e, affordable!) prices directly to consumers, without involving lawyers. In the start-up market, excellent example contracts are being open sourced and made available for free to kick-start standardisation. Like in so many other areas touched by the internet, the value of content is rapidly tending to zero.
Lawyers have tended to struggle conceptually with the idea that what they do is a process, insisting that they are true artists. At the retail end of the market, where so many developments tend to start, workflow has been used for many years by the more sophisticated mass providers, and this is starting to move up-market.
Of course, what’s also quietly happening in the fringes is that processes are just being absorbed by software. More and more interactions with governing bodies that used to require legal support are going to just happen with web pages and help boxes, with only the exceptional cases requiring intervention by lawyers.
There is a huge potential market of consumers and small businesses that need cost-effective help but aren’t being supported by lawyers. A classic area for this is the court system, where self-help has become increasingly popular. This area is ripe for automation and it is noteworthy that eBay created a process for resolving disputes online that was so successfully that it spun the department out as a separate company. These alternative online dispute resolution forums that don’t require lawyers are going to be applied to more and more potential litigations, where already lawyers are perceived as too expensive to be used for anything but the most valuable or emotional of disputes.
I predict that we are going to see more of this in the contracting world too. Products will emerge that help find common ground between contracting parties and help create fully documented contracts with no need to involve lawyers for the more straightforward deals.
But surely, lawyers say, the provision of wise advice can’t be automated? It is certainly true that IBM’s Watson is a long way off from being a consiglieri. However, even here we are seeing automation. Expert systems are becoming increasingly widely used to walk users through a series of questions and then give them expert legal opinions based on the answers. This is not new – Richard Susskind, a forefather of legal technology, was writing systems like this in the eighties.
Beyond the expert system, there are also predictive systems that are using big data-style analysis to start to provide opinions far more grounded in reality than your average “expert”. Great examples that are now being seen are models that predict the outcomes of court cases (systems are showing better predictions of US Supreme Court decisions than experts) and a model in use in the US that can predict with remarkable accuracy the likely outcome of settlement discussions after just a couple of rounds of counter-offers. What is particularly interesting about this system is it can allow the user to try different pathways to model the optimal outcome.
This last example illustrates nicely a theme that is running through all these different technologies. They can either assist experts in performing better (the bionic lawyer model) or they can simply replace an activity that was previously carried out by a lawyer. Either way, we will need fewer lawyers. Meanwhile, most lawyers are completely unaware of most of these advances, and like the frog in the pan of water, are not paying any attention to the rising temperature. The speed at which the law is being automated is therefore not going to be driven by your average lawyer – expect a lot of dazed looks when they finally notice what is going on around them.
To read other sections of our article ‘Age of the Robots’, from Outsource #35 (Spring 2014), click here.
About the Author