With the recent news that the artificial intelligence tool GPT-4 passed the bar exam with flying colors, it is easy to imagine a would-be Paul Revere shouting this at a bar association meeting near you.
The GPT-4 news already is generating a range of reactions, from the head in the sand view that AI can never replace anything that lawyers do to fantastical ideas of “robot lawyers” as a magic solution to access to justice. While neither of these extremes is credible or helpful, these latest AI developments should cause us to take a step back and ponder some important questions.
What does it mean to be a lawyer? Once we consider that, how can we possibly continue to justify the bar exam as the measure of basic lawyer competency? And what does all this mean for access to justice?
Good Lawyers are Not Robots
In my 2022 Resolution for the Legal Profession post, I posed a question for all of us to consider that is worth revisiting here:
Take a few moments to consider what makes a good lawyer in your experience.
(Cue the Jeopardy music here).
Okay, now that you have done that, how much connection do you see between your definition and what is tested on the bar exam? I am going to guess it’s not much and maybe none at all.
On a related note, how much connection do you see between your definition of a good lawyer and what the AI chatbot is doing?
In my own non-expert definition, good lawyers provide good counsel and top-notch services to clients, take our roles as trustees of the legal system to heart, and understand and connect with the communities we serve.
Those qualities really go to the core of who we are as humans, and while AI and other technology can help us be better and more efficient in those roles, it is not going to replace what good lawyers do for their clients and for access to justice more broadly.
The Bar Exam is on Thin Ice
Assuming you agree that we’re more than the sum of our knowledge that AI already is starting to match, this should be the tipping point for the bar exam as we know it today. We should still evaluate lawyer competence, but in a way that better tracks what we actually do as lawyers and our readiness to practice law.
I challenge anyone to tell me that our current version of the bar exam is an accurate measure of what makes a good lawyer by any definition. Because if that is what we really think proves basic competency to practice, we should start making way for the androids.
As I have noted in previous posts, experts looking at these issues already have given us a head start. For example, Jordan Furlong, a renowned legal sector analyst, shared a compelling vision of an entirely new framework for defining and developing good lawyers in two posts earlier this year, Radical roads to reform lawyer formation and Defining lawyer competence. The Institute for the Advancement of the American Legal Profession (IAALS) also has been doing good work on this subject, highlighted by their recent report, Building a Better Bar.
Another Sign that the Billable Hour Should Join the Bar Exam in the History Bin
Technology already was making it possible for lawyers to do better work faster, the real value most clients are looking for but at odds with a pricing mechanism that rewards lawyers for spending more time. Many lawyers have already seen the light and moved to other pricing strategies that focus on value not time.
Too many lawyers serving the consumer market have yet to see that light, and GPT-4 and its progeny are going to supercharge the ability to do things more efficiently and make the time-based billing model increasingly untenable. While we don’t need to worry about being replaced by robot lawyers, “bionic” lawyers who effectively pair AI and technology with their uniquely human lawyering skills are going to have the upper hand.
The Need to Modernize the Way We Regulate the Profession Just Got Even Clearer
For lawyers to effectively utilize these new tools and compete with AI-based legal products, and for the public to have any practical way to know if these tools are reliable and good legal solutions for them, our current regulatory structure for the profession is not going to cut it. We already were way overdue to modernize the way we regulate the profession and delivery of legal services, and that is only going to become clearer with GPT-4 and what is sure to follow on the AI front.
As I have said before, the question is not whether technology-based legal solutions are a good idea for everyone in every legal context. They clearly are not, but these tools are good solutions for many legal issues, people already are using them in droves, and that has only increased during the pandemic.
The real questions are:
- Are we going to do anything to help people determine whether these are reliable products and reasonable solutions for their problems?
- Are we going to enable lawyers to be a part of it?
AI Definitely has a place in Improving Access to Justice
Just because we’re not going to have robot lawyers does not mean there are not huge possibilities for AI as a self-help resource for people, particularly for the many less complex and lower-stakes legal issues people face in the legal system. It can make a huge difference as well in helping people navigate the still overly complicated procedural requirements of our system.
I’ll take a closer look at some of that potential in a future post, but for now here are two examples that are good food for thought on this front.
And second, while the DoNotPay guy definitely stepped in it with his robot lawyer ideas, he has come up with a lot of innovative solutions that can be very effective for lower-stakes consumer disputes, where he says he is going to focus the company’s efforts going forward.
We are only in the Second Inning
It is still very early in the game as we think about the potential for AI and what it will mean for our profession and access to justice. Even so, we ignore the issues noted above at our peril. We can help write the script for a better future for our profession and access to justice if we get out in front of these issues now.