The world stands on the precipice of a transformation powered by artificial intelligence (AI). Anyone who has used ChatGPT, since it was released in November 2022, knows the power of AI to free us from monotonous work and allow us to better direct our attention elsewhere. While we currently are witnessing early applications, AI has a long way to go. GPT-3.5, which many of us have used through ChatGPT, has 175 billion parameters (essentially, what the AI “learns”) — Wu Dao 2.0 has 1.75 trillion, and GPT-4 is estimated to be in the same ballpark.
But the point of this article isn’t to gawk at numbers. The point is to talk about how the technologies enabled by AI are going to change the landscape of the justice system.
Natural Language Processing (NLP, not to be confused with neuro-linguistic programming) focuses on enabling computers to understand, interpret and generate human language — essentially, to translate human language into data, and data into human language. What’s so interesting about this, from a legal perspective, is that law already is a set of rules expressed in language. Computers, as it happens, tend to have been very good, for a very long time, at producing results of certain rule sets (one could argue that’s all they do).
Current applications of NLP tend to be focused on document analysis, language translation and search. Harnessing NLP means more than paving the cow path — in an intelligent system, NLP can help decipher facts, explain law and frame issues. Advanced NLP systems will empower litigants and legal professionals to interact with vast amounts of data quickly and efficiently, and express themselves effectively.
AI creates significant scope for more efficient, cost-effective and accessible dispute resolution. Online dispute resolution (ODR) is more likely to become accessible and easily administrated, and legal information democratized. There will no longer be a requirement to attend a library for legal research or to hire a lawyer to help formulate factual arguments or provide an understanding of the law. Decision-makers may find themselves more effectively able to free themselves from the constraints of human language: with the assistance of AI, evaluating arguments and writing decisions on the basis of their internal logical structure, and the logic of external legal precedent, free from the flaws and nuances of an imperfect language.
In the same vein, legislation may cease to be written in complicated legal language, susceptible to ambiguity, contradiction and interpretation that does not align with the original intent. Instead of being written, legislation may be “coded,” to be translated into human language as needed. Or equivalently, proposed legislation may be translated from human language into data, to be checked for absolute clarity, and to ensure the proposed law aligns with prevailing circumstances (e.g., is data-driven, and is the best way to achieve the legislative objective).
Human language has long been the bane of the justice system: an imprecise tool, used to create laws that aim at precision. We have a way to fix that, and as lawyers in British Columbia, we swore an oath to “conduct ourselves truly and with integrity” and to “uphold the rule of law and the rights and freedoms of all persons.”
As we move forward, we must anticipate AI being invaluable in removing many of the obstacles to access to justice that have plagued our societies, since formal justice systems began. The law will become readily available, in plain language, from anywhere in the world. You will no longer need a 3-year degree, and years of practice as a lawyer, to understand a small portion of the law that applies to you. To participate in the justice system, you will no longer need to be fluent in English, able-bodied, sighted, able to hear or able to read.
That’s a future worth fighting for.