Robot lawyers? Maybe not, but don't discount AI in legal field yet, experts say

Editor's note: This is part of a KSL.com series looking at the rise of artificial intelligence technology tools such as ChatGPT, the opportunities and risks they pose and what impacts they could have on various aspects of our daily lives.

SALT LAKE CITY — Imagine you're in court contesting a traffic ticket. You put on a pair of smart glasses that records the proceedings. And instead of a lawyer giving you counsel, an artificial intelligence program tells you through an earpiece how to respond to the judge.

It's not quite a reality — this very scenario was set to play out in a California courtroom in February, until the British man behind the idea began receiving angry letters, NPR reported. Joshua Browder, the CEO of the New York-based startup DoNotPay, told NPR multiple states sent him threats of possible prosecution and jail time, with one state bar official even saying the unauthorized practice of law is a misdemeanor in some states punishable with up to six months in county jail.

Though "robot lawyers" may not be a reality anytime soon, AI already has a role in Utah's legal field.

In August 2020, the Utah Supreme Court issued something called Standing Order No. 15, which launched a legal regulatory sandbox. The order established a pilot sandbox to assist with monitoring the practice of law by nontraditional legal service providers or by traditional providers offering nontraditional legal services.

"A regulatory sandbox is a policy tool through which new models or services can be offered and tested to assess marketability and impact and inform future policymaking," according to the Office of Legal Services Innovation. "In the sandbox, regulations can be relaxed, data gathered and policy improved."

The sandbox tool was first put to use in the financial sector when it was grappling with tech advances that didn't fit traditional models, such as cryptocurrency. In the legal field sandbox, corporations might employ Utah-licensed lawyers; law firms might enter business partnerships with nonlaw entities; or companies might use tech providers to practice law.

Artificial intelligence legal services

Forty-eight companies are currently doing business in Utah's legal regulatory sandbox, offering services from immigration assistance to personal injury claims. So many companies have applied to the sandbox that the Office of Legal Services Innovation has temporarily stopped accepting new applications.

For some of the companies currently doing business, the sandbox is an opportunity to implement AI into their services. Hello Divorce, for example, uses software to help customers through divorces for as little as $100 and never more than $5,000, according to its website.

Or Jordanelle Block uses software to provide legal services for real estate transactions.

Automated contract review company LawGeex is based in Israel but has an office in Lehi. The company uses AI to review legal documents based on predefined policies — "just like an experienced attorney, but with enhanced speed and accuracy," according to its website. A human team then makes sure everything is sound before returning contracts to clients.

The company serves mostly large law firms and the general counsel of large enterprise businesses. But about a year ago, LawGeex created a new brand: Superlegal, which serves smaller businesses that may not have in-house lawyers.

Ilana Waxman, a marketing manager for Superlegal, said Superlegal is currently developing processes that will allow customers to create contracts in addition to having existing contracts reviewed.

"It's taking contracts off your desk," she said. "It's also accelerating deals so that you can get them done faster and better ... so that businesses can make money at the end of the day."

Christopher Lewis, head of U.S. legal operations for LawGeex, said the company's AI program can access an extensive database of prior contracts, helping it understand industry standards for various types of transactions. The AI is also connected to an algorithm allowing step-by-step analysis as contracts are reviewed, he said.

"Attorneys that rely solely on their own knowledge of contracts tend to miss a certain amount of routine elements," Lewis said, "whereas (AI) enhances that and creates a much higher accuracy rate."

That said, Lewis doesn't foresee a future in which AI completely replaces human lawyers. Each side is good at different things, he said. While a human can't store millions of contracts in their brain, AI sometimes struggles with simple commands or with using natural language.

"We have found that it's the marriage between artificial and human intelligence that really creates an exceptional product," Lewis said. "I don't think (AI) is going to take over the planet. I don't think it's going to unemploy lawyers. ... It's a tool and it's an exciting tool, but it's not a threat to humanity."

'Explosive increase' in AI

Contract work isn't the only way AI is being applied to the legal field. Paul Cassell, a criminal law professor at the University of Utah's law school, said the judicial system sometimes uses algorithms to help determine bail amounts or release conditions. While judges have the final say, the algorithms can provide guidance as decisions are made.

Cassell said while these algorithms aren't widely used in the judicial system right now, he believes the law industry is standing on the cusp of an "explosive increase" in artificial intelligence usage.

That usage might not always be obvious, he added, such as when lawyers use chatbots to draft briefs. But at the end of the day, "that brief is still going to be filed in the name of the lawyer ... as long as the lawyer is there and has to take ultimate responsibility, then I think things are going to work out."

AI can also make legal help more affordable and therefore more accessible, Cassell said. For instance, if chatbots are eventually capable of writing the same quality of briefs as a human, lawyers will no longer have to bill clients for the 10 hours it would have taken them to write the brief themselves.

While AI can potentially make lawyers more efficient, it's also presenting a challenge at law schools, Cassell said.

"We need to be thinking about how this is going to work and training the future generation of lawyers to be able to capitalize on these powerful new tools that are coming. ... Maybe in the future, we'll be giving grades to students and chatbots working together."

Will AI replace attorneys, judges?

Dane Thorley, however, thinks AI-related changes aren't "coming down the pipe quite as quickly as some people feel that (they are)."

Thorley, an associate law professor at Brigham Young University's law school, researches how courtroom rules and procedures impact the behavior of judges, attorneys and other courtroom parties.

Once, while researching how bail decisions are made, Thorley said he learned about a number of companies that created predictive algorithms around 10 years ago. These algorithms used a defendant's information — where they lived, family and religious background, education level, racial and ethnic categories, prior crimes and more — to predict the likelihood of them committing another crime if they were released.

But none of these companies publicly published their algorithm information, Thorley said, meaning no one could be sure how the algorithm was weighing all of the factors.

"To the extent that we think that the courts are an extension of government, and we think the government should be governed by democratic principles, then we have a transparency problem, because we don't know what the government is making their decisions based off of," Thorley said. "(And it's) not just (that) we don't know what information is being fed to (the algorithms), but the information fed to it may be biased."

Thorley added artificial intelligence usage isn't going to happen all at once — courtrooms aren't going to suddenly start replacing judges with AI or allowing people to use AI attorneys.

But if AI ever did completely replace human judges and attorneys, Thorley said he'd be concerned about equal access to justice. An AI algorithm is only as useful as the information it's loaded with, he said. And since that information is likely to be from past cases, people of color will naturally be underrepresented, particularly in places like Utah, he added.

"So we have a greater likelihood that there will be problematic disparities (in) racial and ethnic categories because the information that's being fed into the AI algorithm is going to be less complete for certain groups of people," he said.

Correction: A previous version stated that Ilana Waxman is a marketing manager for LawGeex. She is a marketing manager for Superlegal, a brand created by LawGeex.

Related topics