Ghosts in the legal machine: How artificial intelligence is transforming the field of law
GRAND FORKS, N.D. – For a reporter, you know you’ve discovered a good trend when companies say it’s so hot, they can’t talk to you about it.
In December, that happened to this reporter on the topic of artificial intelligence and the law.
Yes, we’re working with AI programs to improve our practice, said a spokesperson for a Minnesota-based law firm.
But the attorneys most familiar with the technology “are very sensitive about talking about it,” the spokesperson said.
“It’s a very competitive environment around innovation and technology right now, so they’re just concerned about letting the cat out of the bag about what we’re doing.”
Understood. But what is this trend that’s so sensitive and competitive, it can’t even be discussed?
The issue is the trend’s power. Because the coming of AI means “the day-to-day work of lawyering is going to change dramatically over the next five or six years,” said Allen Blair, professor at Mitchell Hamline School of Law in St. Paul.
“One of the traditional things about law is that it is slow to change, and regarding technology, law has been blissfully slow.” But “that’s no longer going to be the case,” Blair said, as the technological whirlwinds that have disrupted other fields start moving in the direction of law.
Francis Shen, associate professor at the University of Minnesota Law School, agreed. For two years, Shen has taught one of the region’s comparatively few courses on artificial intelligence and the law. “And in my seminar, both years, students have concluded that big changes are coming, and that law had better prepare for them,” he said.
AI and the law firm
Artificial intelligence’s impact on law firms is being felt in two broad realms: internally, in how the firms do business; and externally, in how firms are tackling the legal issues raised by AI in other industries.
Let’s consider the firms’ inner workings first.
Traditionally, computers are programmed to do logic, said Michael McGinniss, associate professor of law at the University of North Dakota School of Law in Grand Forks, N.D.
But AI adds experience. “You teach the machine what to look for, and then it gets better at it,” he said.
“You’re actually teaching it to simulate what a human would do in a similar task. That’s the core insight that makes AI different from traditional computer work.”
So, law firms now are using AI to improve many aspects of legal work, such as discovery.
“I was just consulting on a case where we had 2½ million documents,” said Blair of Mitchell Hamline.
“When I first started practicing, you would hire a company that would put people to work looking for key words. Dozens and dozens of temp employees, reading documents all day.”
Today, e-discovery uses AI systems that get better with every round of searching. Ultimately, they’re scanning the documents not only for words and phrases, but concepts and ideas.
“It’s faster, it’s more accurate, and it’s able to get through a volume of material that humans just can’t,” Blair said.
These days, “most law firms, if they have even moderate-sized cases, likely are using this kind of artificial intelligence for discovery.”
A newer function called predictive analytics is moving so fast that the latest development was announced Nov. 29. The following is from a press release from that date describing Context, a product by LexisNexis, a company that provides legal research:
“Context is the first and only case-law language analytics solution in the industry,” the press release reports.
“The solution analyzes tens of millions of court documents and extracts specific language that will resonate with a particular judge … to help you draft winning briefs or successfully argue motions. …
“Said Jeff Pfeifer, vice president, product management at LexisNexis: ‘When attorneys know what specific language a judge regularly uses ... they are empowered to argue more persuasively and effectively on behalf of their clients. … Only Context delivers this level of insight, right from the very start of the litigation process.’”
Of course, lawyers always have asked each other, What are our chances with this judge?
Now, computers are being taught to estimate those odds. And in time, they’ll likely make better guesses than humans can, thanks to AI.
“Which raises some serious due-process arguments,” Blair said.
After all, are courtroom decisions governed by law, or by the biases of an individual judge? If it’s the latter, then “that may no longer be the kind of fairness we’ve come to expect the law to deliver,” he said.
“In any event, it’s going to impact the way judges and courts and resolution works.”
AI and clients
Then there are the questions that AI is raising in other industries – questions that lawyers are being hired to help answer.
“A big issue in law is that humans have always been in the loop,” said Shen of Minnesota. “Always, always.”
There is a doctor who makes a diagnosis. There is a person behind the wheel of a car.
For decades, doctors and drivers have been helped by technology. But what happens when the technology itself is diagnosing patients or driving the car?
In other words, who’ll be held responsible for a crash involving a self-driving car?
“The law was based on governing human error and human mistakes,” Shen said.
“Now, it’ll have to be re-engineered to account for algorithmic mistakes. That’s what lawyers will be paid to do.”
A fascinating case in Wisconsin shows another twist on AI.
Judges consider lots of information when making sentencing decisions. But in the case of Eric Loomis, who was sentenced by a Wisconsin judge, one key report had been generated by a software program’s algorithm.
The algorithm found Loomis highly likely to reoffend. But when Loomis asked to inspect it, his request was denied, because the company insisted that its algorithm is proprietary.
“In other words, it’s in a black box, and we can’t tell you how it works,” Shen said. “We can just tell you what the result was, which is that you’re not getting out of jail.
“Clearly, that raises some deep questions, and I think we’re going to see more of that.”
One last example, this one involving contracts.
“I teach something called the ‘playground logic’ of exchange,” Blair said.
“If I hand you a candy bar, and at the exact same time you hand me a dollar, we don’t need a contract.” That’s because the playground exchange is completed instantly and doesn’t require trust, which is what contracts enforce.
“Today, blockchain technology allows for that sort of playground logic in the renting of hotel rooms, the buying and selling of cars, any and all kinds of everyday exchanges,” he said. And AI promises to make “smart contracts” smarter still.
“In effect, this could make contract law irrelevant,” thereby upending an entire field of law, Blair said.
A profession ‘rooted in data’
The net result of all of the above is clear, the professors agreed.
“I think that the practice of law is about to become a lot more exciting,” Blair said.
Gone are the days when a lawyer could look forward to a quiet career of filling out forms. Instead, “it’s going to become a much more heady profession, one rooted in data.”
“I tell my students they’ll have to know at least enough about statistics and basic math to understand what happens with predictive coding,” Blair said.
Shen agreed. And that’s a challenge, he noted. After all, “a lot of law students go to law school because they’re not good in math.”
But they’ll have to adjust, Shen suggested – because if they don’t, the changing profession is likely to leave them behind.
Editor, Prairie Business