Why Some Lawyers Should Be Expert Coders

An example of law as code, implemented in the Ergo programming language from Coherent Knowledge.

Another round of “should lawyers learn to code” started up in the Twittersphere this week. The arguments against it are not new.

  • “Lawyers should learn to work with coders, not be coders.”
  • “Specialization of skills is better than a person who tries to know everything.”
  • “Lawyers don’t need to learn to code, they need to learn to use Microsoft Word properly.”

The defenders pipe up and say:

  • “There is a difference between teaching someone to code, and making them a developer. We are encouraging literacy, not expertise.”
  • “Lawyers need to understand what algorithmic bias is, and where it comes from, in order to help their clients deal with it.”
  • “Learning to code doesn’t stop you from learning to use Word, and actually makes it easier.”

My view is very different.

I have recently come to realize that for people who don’t code, “coding” sounds like something different than it is. It sounds like a craft. It sounds like pottery, or knitting. It sounds hard to do, and designed to create a specific type of object — a computer program.

A computer program is a useful thing, as is a pot, or a sweater, and some of them are even useful to lawyers. But it is not the job of a lawyer to make good sweaters. So coding seems a distraction from our job.

Here’s the thing: for people who have done coding, coding does not seem like a skill. Coding seems like a language used to communicate with the most efficient helpers available. Or, as David Colarusso said on twitter recently, coding is “expressing rules in an unambiguous manner.”

Put that way, of course that is something lawyers should learn.

But I would go further than David, who makes the “literacy” argument. I say we need expert lawyer-coders. Not every lawyer needs to be one, but we need some, and we need them now.

A lawyer who cannot code cannot audit the correctness of software that provides automated legal services. And absent competent automated legal services, we have no hope of addressing the access to justice problem.

If a coder was attempting to address access to justice with a software tool, and came to the conclusion that they needed to be sure the tool was legally correct, and decided to study law, no one would balk.

Why the same observer might find coders learning the law inoffensive but insist lawyers shouldn’t learn to code is a very interesting question. Perhaps “interesting” isn’t the right word. “Worrisome,” maybe.

The starting point for this argument doesn’t need to be access to justice. It might be algorithmic bias, or smart contracts, or what have you. For various reasons, there are legal services which we clearly need, which clearly cannot be effectively provided by a lawyer who cannot read and write, at an expert level, the language in which the relevant rules have been expressed.

We cannot get by in a world that operates in one language and lawyers who only speak another. Lawyers need to learn to speak the language of the law at an expert level, and the language of the law is increasingly the language of computers.

But, a lot more could be done to make it easier for lawyers to learn to speak code. I’ll talk more about that next time.

Jason Morris is a lawyer at RoundTableLaw.ca, an LLM student in Computational Law at the University of Alberta, and a 2018/2019 ABA Innovation Fellow working on an open source tool for analogical legal reasoning in expert systems. His TEDxUAlberta talk on how code can increase access to justice is available here. He is @RoundTableLaw on Twitter.

Lawyer, Round Table Law; 2018/2019 ABA Innovation Fellow; Sessional Instructor, University of Alberta; Computational Law (Symbolic AI) Researcher, CCLaw @ SMU

Lawyer, Round Table Law; 2018/2019 ABA Innovation Fellow; Sessional Instructor, University of Alberta; Computational Law (Symbolic AI) Researcher, CCLaw @ SMU