“Rules as Code” Extended Abstract to appear in ICAIL ’21 Proceedings

The International Conference on Artificial Intelligence and Law is a renowned international academic conference on law and artificial intelligence. It is held every two years.

In 2019, I had the pleasure of attending in order to present a demonstration of work that was part of my LLM in Computational Law at the University of Alberta.

I’m happy to share that I will be “attending” again this year. ICAIL ’21 will be publishing an extended abstract of mine, which is one level above a demonstration. I suppose that in ’23 I should aim for a short paper, and in ’25 for a full paper to hit for the cycle.

The abstract describes an experiment in which we encoded a section of legislation in the s(CASP) programming language. Our experiment was designed to mimic a Rules as Code co-drafting process, and reveal the strengths and weaknesses of s(CASP) as a tool for improving legislative drafting.

Not only did s(CASP) prove itself remarkably useful as a tool for Rules as Code, but quite unexpectedly we actually found an obscure legislative drafting error, coded a proposed amendment, and proved that the amendment resolved the legislative bug.

People have been using logic programming languages to encode statutory texts since Kowalski did it in the ’80s. And attempts at improving legislative drafting through formalisms have been around longer than that.

But as far as I know this is the first time anyone has described the use of symbolic artificial intelligence to find legislative bugs in a law while it is being drafted, which makes it the first published work on “Rules as Code.”