Could AI Open Up the Legal Profession? Not Anytime Soon.

Legal jobs top the list of those highly exposed to AI, but that doesn’t mean hiring and the path into high-paying roles will be changing anytime soon.

Last year, researchers in New Zealand ran an experiment. They pitted junior and senior lawyers against large language models in a race to review contracts. 

The findings were stark. The technology was just as accurate as the junior lawyers and took mere minutes. The cost per contract was in some cases just pennies. 

The study wasn’t the only one to identify the legal industry as one ripe for disruption. As technology has improved at completing verbal tasks, the daily work of junior lawyers and paralegals has moved into the crosshairs of AI. 

The big idea: Experts say that a revolution in the labor structure of the legal industry—one in which workers without a pricey J.D. can take on the work of lawyers—may be conceivably possible, but is certainly not around the corner. 

“I wouldn’t expect groundbreaking changes should be happening in the next six months to one year,” says Manav Raj, a professor of management at the University of Pennsylvania. “This is going to be a longer pathway.”

It’s a pathway that many professions are on—making this a wait and see moment for millions of workers, even as companies are experimenting with AI and mapping its possibilities.

Noah Weisman, founder of the legal machine learning company Kira, says he doesn’t think advancements in technology will reduce the number of lawyers. Kira, which reviews contracts, hasn’t resulted in any changes in hiring, to his knowledge. 

New efficiencies from AI may simply increase the amount of legal work to go around, Weisman says. In economics, this phenomenon is called the Jevons paradox. Broadly speaking, it describes a situation in which efficiencies induce demand, rather than weaken it. 

Namesake economist William Jevons observed this in 19th-century coal, but Weisman points out the same effect in modern refrigeration. As refrigerators have become more energy efficient and less expensive to run, they have also become more prevalent. 

With lawyers, expanded efficiency might look like going deeper on a task. If attorneys typically review just a percentage of a company’s biggest contracts during an acquisition, AI tools could examine the remainder. Instead of cutting down on the amount of legal work, the technology only increases it. In theory, that sort of development could also expand access to legal services to people who historically haven’t been able to afford it. 

“As AI unlocks that greater efficiency, you have real opportunities for creative people to expand the amount of legal work that is done,” Weisman says. 

It’s certainly possible for that future to include people without a legal background doing legal work, Weisman says. But in his speculative scenario, those are workers in foreign countries where labor is cheap, not upwardly mobile Americans

To be sure, law is a highly regulated industry. In order to practice law, one must pass the bar exam. While it’s possible to pass without a law degree in some states, it’s extremely rare. That puts some limits on what non-lawyers are allowed to do. Paralegals, for example, are typically barred from giving formal legal advice. 

However, what sort of activities count as practice of the law has never been clearly defined. 

“There is a lot of ambiguity about what it is that lawyers can do that people who are not licensed to practice can’t,” says Andrew Perlman, dean and professor of law at Suffolk University.

Raj’s research on the occupations that have the most exposure to AI and large language models places legal professions near the top of the list. That’s likely, he says, because legal professions are heavily involved in verbal skills, such as reading and writing. 

But exposure to AI, as Raj and his coauthors see it, doesn’t mean that those jobs will be replaced by AI. It also doesn’t necessarily mean there could be large productivity gains from AI. 

“We suspect, among these occupations some will see kind of the upside and maybe there will be some downside in some cases,” Raj says. 

Large-scale data processing, like looking through large sets of documents, is something AI can be useful for. Greater deployment of AI in those areas might reduce the need for paralegals, Raj says, or change the sort of work that they are doing. 

Paralegals may need to become subject matter experts, says Mark Cohen, founder of the consulting company Legal Mosaic. In fact, more and more subject matter experts without law degrees may be working in the legal industry in the future. The industry, with greater capacity, may start becoming more proactive at avoiding litigation. 

But how soon any transition happens, and to what extent it happens at all, still remains unclear.  

“There are going to be lots of jobs for lawyers, but they’re going to be very new jobs,” Cohen says. “The legal industry, left to its own devices, has demonstrated time and time again its resistance to change.”

Related Posts

American College Students Lag in Using AI

About 38% of college students in the United States now use AI tools at least monthly—but only 10% use them at least weekly, a far lower rate than in other developed countries. Only British students lag further when it comes to familiarity with AI tools, according to a new survey of 10 large countries by ... Read more
Download the Work Shift Guide to Understanding New Collar Apprenticeships