Robots are doing some of what lawyers do, and they will do more, but the world is not done with lawyers yet - at least for the foreseeable future.

The legal business is no more immune from automation than any other industry, with the driver being cost. Clients demand more and more for less, and our increasingly networked world means that clients can get services from anywhere; relationships are always important, but cost is never irrelevant. The successful lawyers will be those who adapt to this dynamic environment, and there are plenty of developers out there who see the opportunity presented, and are hard at work tackling the challenge.

You don’t automate for the sake of it. There are several factors at play in relation to any given task or industry (McKinsey, July 2016). First, it must be technically possible to replace a human doing a given job with a machine doing the same job. Second, creating that machine has to be worth the cost of doing so, and the cost of the replaced labour less than the cost of the machine. So scale is very relevant here. Third, the human factor has to be expendable, and so not essential to the successful execution of task in question.

The job of a lawyer is multi-faceted, and not all are suitable for automation – or at least yet. Automation is most obvious for tasks that entail data entry and /or data analysis, carried out in a repetitive and predictive environment. Robots are already employed in some of these areas, or their use is in development.

For example, ROSS is a robotic paralegal “employed” by a large US firm to help in its bankruptcy practice. Other ‘lawbots’ are in operation, or being trained, to conduct due diligence (on property transactions, contract analysis as part of M&A due diligence, “know your client” information), prepare and check basic loan documentation, undertake transaction workflow management. There is also a programme in operation in some states of the US which makes sentencing recommendations to judges, based on algorithmic assessments as to likely re-offending. Another app is in operation in the UK, helping people get off their parking tickets. There are also businesses offering software-driven legal precedent systems, where first drafts of standard documents are created based on answers provided by the user to a set of questions. (Delivery of legal services on-line is not new – email is standard, and there are legal businesses in operation now whose services are solely delivered on-line, but in ‘commoditised’ areas and with humans still working behind the digital scenes (as it were), and who are on call should a human intervention be needed.)

The common thread (so far) with the current robot roll-out within the legal industry is that the work being undertaken by the lawbots is confined to a certain area – bounded in some way such that the machine can learn the rules, and so come up with reliable answers, more cheaply than a human. The more that the machine does within that area the better it becomes.

The robotic capability in the legal industry will become more prevalent as the extent of data analysis, computing power and machine learning grows, and – importantly – as the cost of the technology drops. If a firm’s business is more or less solely focused on large scale M&A deals, or large scale litigation, there is economic sense in investing in robotic technology to eliminate the need for an army of humans conducting basic contract review work as part of a due diligence exercise, or the equivalent in pre-trial discovery. But the cost is simply not worth it for businesses who aren’t as specialised, and who don’t have the scale. However, given that DD and discovery is conducted in more or less the same way world over, when the automated ability is cheap enough and widely available, its use and acceptance will become as commonplace as email, and with predictable results on the numbers of humans in law firms who would otherwise be employed to do this stuff.

The final question though, ahead of a complete technological takeover of the business of law, is whether the human aspect is still valuable. The answer is probably ‘yes’ - Robo-Harvey won't cut it (yet). 

While legal services are still being provided to humans, with the potential to significantly affect lives and livelihoods, there is a need for a fellow human to be involved in the delivery of the key elements of that service.  So far, the human is needed to judge how best to present information, and to persuade – and, crucially, be trusted by the person receiving the information.  The human practice of law will remain to deliver those aspects of the job that needing the human overlay: delivering of complex advice, advising on strategy, nuanced / specific drafting, negotiation, trial advocacy, judgements that require a determination of what is “right” or “fair”, involving as it does a balancing of moral, ethical and black letter factors. This position will change completely however as and when we reach the point of (acceptable) singularity.  When an algorithm is created that enables a machine to provide the output that would otherwise be human, and that ensure the recipient trusts what is produced.

For now though, quite apart from the sheer entertainment value of Harvey Specter et al (and their ability to do an IPO in 4 days, and a merger in 3!), he and his team are here to stay for the foreseeable future.  We still need Harvey to be clever, cunning and charming. Harvey still needs Mike to do all the persuading of reluctant humans that he does (although parts of Mike’s life will become easier). And we’ll need both of them to brainstorm that left-field strategy to impress Vanessa, and defeat Lewis. 

Most importantly though, we need Harvey to offer the wisdom gleaned from his experience, and for Mike to get similar experience and so eventually become wise for when Harvey has gone.