Amid the hype and headlines about robots taking lawyers' jobs, the depiction of artificial intelligence in relation to humans is often pitched in an adversarial way. When viewed through this prism – human vs machine – it's a zero-sum game; one side wins, the other loses. However, as I often find myself saying, robots are not usually replacing lawyers, they are assisting them – and a lawyer who uses technology wisely will beat a lawyer alone or technology alone.

It's not just within the law that we find this tendency to pit people against technology to see if one can 'beat' the other. You are probably familiar with some of the more famous of these battles, staged by technology companies to showcase their wares (and sell them). Big Blue beating chess grandmaster Gary Kasparov, and Watson beating the long-standing Jeopardy champion are the two that stand out. These were huge achievements and demonstrate how far technology has come, so should not be underestimated. But they are also not necessarily a fair fight – armies of engineers and huge investments of technology pitted against a single human is an interesting spectacle no doubt, but what does it prove? That technology, if sufficiently advanced, can outperform a single human in areas with fixed rules. Not entirely surprising. But what about humans grouped together against the computer programme (which, after all, is the product of the work of hundreds of humans), or humans working with the machine?

Usually, when we are talking about robots or advanced tech, we're talking about learning machines – technology that can be trained from the data it is given and can make inferences and decisions based on what it learns. And what this technology is really good at is pattern matching – recognising what it is looking at based on having learned certain characteristics of the domain in question – e.g., dogs, cancer cells, clauses - and identifying where data are similar or different. We can all get a little defensive at the idea of a machine doing something that we consider ourselves skilled at, and the tendency is to expect it to do everything, and to write it off if it can't. The reality is that machines can do a lot of the tedious and time-consuming tasks, freeing up the human actor to do the creative, challenging and rewarding work. But the human needs to meet the tech halfway.

When Kasparov was beaten by Big Blue, he didn't shun technology and retreat back into the world of human vs human chess. He became an advocate for a new type of challenge, where human and machine work together, called Centaur Chess. By letting the machine assess myriad moves for potential problems, the human can focus on strategy and creativity, and working together they can beat any computer program working alone. Interestingly though, he added another element to the equation – that of process. Kasparov's formulation was that a weak human chess player plus a machine with a very good process, can beat a strong human chess player plus the machine but with a weak process. That is, a grand master could be beaten by an intermediate chess player if the latter were using a robust process that incorporates the technology's strengths with their own.

Taking an analogous approach to legal work, the challenge for lawyers is to become proficient at managing the process of using machine learning technology, in order to get the most from it. The human, in this case the lawyer, has to make the effort to understand what the technology can do, so they can learn how to benefit from combining their own skills with those of the machine, and ultimately become a Centaur Lawyer.


Damien Behan

Innovation & Technology Director