AI here, AI there, AI everywhere. That seems to be the trend. But are we willing to cede good lawyer skills to a bot? That seems to be a risk according to a white paper from Thomson Reuters.
There’s a famous quote attributed to the science fiction writer William Gibson: “The future is already here — it’s just not evenly distributed.” The white paper demonstrates this very point: AI is eroding critical thinking skills at an alarming rate. The future will be distributed to those who figure out how to retain and enhance these skills.
The Paper
The white paper amplifies a troubling trend that I have discussed before: AI is eroding lawyers’ critical thinking skills. Reading the paper confirms what many, including me, have feared: “As AI becomes more capable, lawyers risk becoming less so.” Without these critical thinking skills, a lawyer simply cannot exercise analytical skills to identity and define legal problems, much less find solutions.
The paper was written by Valerie McConnell, Thomson Reuters VP of solutions engineering and former litigator and Lance Odegard, Thomson Reuters director of legaltech platform services.
The Current Threat
The findings should scare the hell out of seasoned lawyers:
The headline? Research from the SBS Swiss Business School found significant correlations between AI use and cognitive offloading on the one hand and a lack of critical thinking on the other. Critical thinking down, cognitive offloading up.
McConnell says that “cognitive muscles can atrophy when lawyers become too dependent on automated analysis.” Odegard adds an even more concerning fact: AI is different than previous technologies given its speed and depth. And the fact that it can perform some cognitive tasks creates a greater risk of overreliance on it.
I recently attended a panel discussion of law librarians on the use of AI in their law firms. One telling remark: more experienced lawyers were able to form better prompts because they understood and could better articulate the problem than less experienced ones. And they could quickly determine whether the output was bogus: when it didn’t look or sound quite right. They got these skills through developing a critical way of thinking from seeing patterns and prior experiences. AI short circuits and replaces the pattern-recognition experiences.
The classic example of this is where the AI tool explains a legal concept with certainty but the explanation doesn’t not look right to an experienced lawyer who has dealt with that concept and understands how and why it was developed.
The Accelerated Risks Of Agentic AI
But there’s more danger ahead according to the paper. Agentic AI can perceive its environment, plan and execute complex multistep workflows, make real-time decisions and adapt strategies, and proactively pursue goals, all without human input. This means, according to the paper, that agentic AI could intensify cognitive offloading. In other words, we turn off our brains and let AI do the thinking for us. And as discussed before, we don’t have a clue how it is doing all this.
McConnell and Odegard believe agentic AI creates “unprecedented professional responsibility challenges.” How can lawyers ethically supervise the systems? What levels of competency will we expect and demand from human lawyers? How will lawyers ethically communicate with clients about strategies developed by the “black box”? Lawyers have an ethical duty to explain the risks and benefits of strategic options: how can we do that when those risks and benefits are developed in ways we don’t understand?
I recently wrote about the phenomenon of legal tech companies buying law firms and the danger of a reduced lawyer in the loop. Agentic AI magnifies these dangers significantly.
Do We Need Critical Thinking?
As with any “truism” it’s always useful to pause and reflect whether it’s really a truism: how much will future lawyers even need critical thinking skills when AI can do it for them?
McConnell and Odegard certainly believe that future lawyers will need these skills. They believe that AI cannot replicate these skills, nor can it yet replace the creativity and nuanced understanding of a good human lawyer.
I agree with them on this point. I see it frequently as AI spits out solutions as if handed down from above. And it sticks to its guns even if wrong. The fact that the tools are so easy and quick to use also makes it pretty tempting to just accept what it says without thinking it over. This is especially the case for busy lawyers.
And that’s one reason we are continuing to see hallucinated cases cited in briefs and even judicial opinions.
But what happens when we rely on the bot instead of our own instincts borne out of experience? A few years ago, I trusted the handling of a significant hearing to local counsel. The day before the hearing, I got the feeling after talking to the local counsel that something was not quite right. So, I quickly hopped on a plane and went to the hearing myself. Good thing: the local counsel didn’t show and sent a first-year associate to handle the critical hearing. I doubt a bot would have picked up that nuance.
The Risks For Future Generations
McConnell and Odegard also cite the danger of overreliance on AI to replace these skills will erode younger lawyer development. It may result in lawyers depending too much on AI instead of thinking for themselves. It may result in “lawyers skilled at managing AI but lacking independent strategic thinking.”
I too have discussed this very real problem. Doing what many call scut work as a young lawyer was boring and tedious, but it helped you begin to see patterns that could be helpful later in similar circumstances.
But now we are urged to dump these tasks into a chatbot and forget it. The result in 10 years? Minds full of mush. The old notion of thinking like a lawyer may be replaced by thinking like a bot.
Another danger: the erosion of legal education. According to the paper “students increasingly arrive with diminished critical thinking skills due to pre-law AI exposure while expecting to use AI tools throughout their careers.” If we don’t take steps to disrupt that expectation, we can be sure that these students, when they become lawyers, will continue to use AI tools in exactly the same way.
Can The Risks Be Managed?
To be fair, McConnell and Odegard believe these risks can all be managed by responsible use of existing AI tools. That may be true but as with most technology, some lawyers and legal professionals will figure out how to do this and become future superstars. Many will not. And maybe that’s OK since many legal jobs and work done by humans will be replaced by AI.
Certainly, AI will allow lawyers and legal professionals to do the high-end stuff for which they were trained. But let’s be real here: there is not enough demand for the high-end work to go around. And many lawyers and legal professionals are not that good at it.
The Future: It Won’t Be Evenly Distributed
So, want to prepare for the future? Figure out how to encourage and develop critical thinking skills among your work force in the age of AI. Figure out what to do when the only work to be done is high-end thinking. That means preparing for a law firm that looks very different from today.
Get ready for the future, it’s not going to be evenly distributed.
Stephen Embry is a lawyer, speaker, blogger, and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law.