Column

Let’s Chat About ChatGPT

Three months ago, I had never heard of ChatGPT. Now, a day doesn’t go by when I don’t find myself talking about it. How will it impact teaching? What about exams? And most importantly of all, how will it affect the practice of law and the work of the judiciary – for in the end, that’s what will determine the answers to the first two questions.

The locus classicus when it comes to the adoption of new innovations is a 1962 book by Everett Rogers, a professor of sociology at the Ohio State University, Diffusion of Innovations. In his book, Everett argues that innovations – whether social or technological – typically are rolled out in five stages, according to a sine-shaped curve. There are some (and I am one) who argue that our inherent conservatism as a profession means that Everett’s analysis isn’t applicable in its original form to law. Simply put, there isn’t a widely perceived first mover’s advantage in the legal profession. No managing partner I know wants to be out ahead of the pack. Nevertheless, where Everett’s thesis is applicable to the law is his contention that adoption of innovations is a gradual process. At least until now. The fact is that ChatGPT has taken hold in a way that, even if we had known about it four months ago, none of us who studied how the legal profession operates would have ever believed possible.

The “GPT” in ChatGPT stands for “generative pre-trained transformer. It’s is basically an artificial intelligence tool that can be applied in any number areas, including the law. Ask it a question, and it will scour the internet, and generate an answer in any format you wish. Want it in memo form? Check. A series of bullet points? Ditto. It will even generate something approximating an opinion letter!

Early on – by which I mean a few weeks ago – some lawyers I know scoffed because ChatGPT made some elementary blunders when it was asked questions on points of law. Similarly, there were knowing nods in some circles when it failed one of the US Bar exams. But those who think that way clearly don’t appreciate how quickly computer programming can evolve. For more recently, ChatGPT has not only mastered the Rule Against Perpetutities, but it has passed the Bar. And lest anyone think that our students aren’t ahead of us, I’ve been told by a number that they’re already using it in their assignments. I’d be very surprised if a good many applicants to law school aren’t using it to generate their personal statements. Indeed, I’ve even had a judge tell me that he already uses AI to help him write judgments!

So ChatGPT is here to stay. Not only that, but it’s going to become more and more sophisticated, and hence useful, with each passing month. And what that means is that if it doesn’t do so already, our professional duty of technological competence will soon include being able to use ChatGPT and other AI tools. If this is the case – and it is – my fellow law professors and I need to start incorporating it in how we teach. Adeptness with AI will be no less important a skill for tomorrow’s lawyers than the ability to use the Canadian Abridgement was for us.

What of the practice of law? Are we all destined to be made redundant in a few years? Is the dystopian nightmare on the verge of coming true? The short answer is no – provided we are willing to embrace change. When I was a young lawyer, we all deferred to the lawyer who knew the most law. But it has been a long time since knowing the law was what separated the grain from the chaff. Want to know relevant legal principle in a given area? Forget the Canadian Abridgement, in most areas of the law, you can Google it in a few seconds. No, the key now to being a “trusted advisor” (which is what every lawyer I know aspires to be) is to possess sound judgment, both strategic and tactical acumen and the ability to write prose that sings. If they cultivate those skills, the future can be very bright.

There is no question that the development of ChatGPT and the tools which will inevitably follow it is posing a tremendous challenge to law schools and to bar admission courses. We’re all going to have to recalibrate our professional sights. But there is no reason we can’t do that. And that should give today’s lawyers – and today’s law students – hope.

Ian Holloway is the Dean of Law at the University of Calgary. 

Comments

  1. Melanie Bueckert

    As a member of the Canadian Abridgment’s Editorial Advisory Board, all I will say is, please DON’T ‘forget the Canadian Abridgment’. I agree that we need to be willing to embrace change and that it is acceptable to use Google in the legal research process, but topically-indexed sources curated by knowledgeable human beings still play a very important role in efficiently and effectively resolving legal problems.

  2. Heather Hui-Litwin

    Thank you for this article, Dean Holloway! It inspired me to check it out! Here is a sample demo of what ChatGPT did when I asked it to draft a Statement of claim. I also asked it what rules apply. https://youtu.be/IaRstaFpRTM. You can definitely see its limitations but it is pretty neat!

  3. Thanks, Melanie. You’re absolutely right – expertise still counts. Judgement, too, which comes from expertise. (And for what it’s worth, I think that the old style of research, including the Abridgement with all its quirks, helped train my mind to be more questioning and analytical.

    Thanks to you, too, Heather!

  4. Melanie Bueckert

    For further information regarding ChatGPT’s evolving capabilities, check out this article: https://www.visualcapitalist.com/how-smart-is-chatgpt/. For example, an earlier version of ChatGPT scored in the 40th percentile for the LSAT; the latest edition scored in the 88th percentile.