Experts say artificial intelligence can be useful in legal work, but caution should be key

It’s no less than a legal obligation for any lawyer to assure that whatever he or she expresses to the court is truly accurate.

And with artificial intelligence (AI) slowly but increasingly weaving its way into the courts and legal landscape, it’s never been more imperative for those filing briefs or conducting research with the help of AI to assure the accuracy and legitimacy of all documents filed.

Some may argue AI creates a more robust or sophisticated way of producing court filings. But with increasing potential for fake sources, not all who practice law are fully jumping on the AI bandwagon.

Georgia-based personal injury attorney Darl Champion lives up to his name when it comes to reviewing motions, drafting rulings and reading case filings. Relying on his “old-school” training, rather than depending solely on AI for assistance, he believes the technology is far better suited as a tool rather than a definitive source.

“When I started back in the day, we had Westlaw and Lexis for our electronic database searches,” Champion said. “You still had a lot of people using books, going to the case reports, finding the case and researching issues. Call it Legal Research 101.”

A graduate of the Mercer University School of Law in Macon, Georgia, Champion spent two years clerking for a U.S. District Court judge in the Peach State prior to opening his highly successful practice.

Even as a spry 43-year-old, Champion finds it troubling that so many “young attorneys” are relying on AI for legal help.
“Today there are so many technology tools available and because so many lawyers can and do practice for so long in their careers, there are a lot of interesting generational differences,” he said. “Even at 43, I didn’t grow up with social media or smartphones.”

His life experiences tell him that different times mean different things to different people.

“People coming out of law school today probably have a much different view of AI than many older folks who are still practicing law. And I don’t want to sound like the old, ‘get off my lawn, curmudgeon type of guy,’ but I think there are a lot of inherent problems with AI.”

For starters, there is the issue of what is classified as “AI hallucinations,” information considered fabricated or generated by the technology. Plus, many legal professionals, Champion included, have argued that human biases can produce biased AI results which can lead to serious consequences.

“Once you cede authority to a computer program, there is also a greater tendency for people to rely on technology tools,” he said. “I feel it encourages complacency and mental laziness that can be harmful to a client.”

Practicing personal injury law since 2009, Champion adamantly stresses the need for any legal practitioner to properly vet all documents destined for a courtroom, or risk being seriously disciplined.

“I’m hearing more and more stories of lawyers getting sanctioned for filing documents with the courts containing completely made-up cases cited in them,” he said. “In fact, one lawyer actually used the excuse of why he shouldn’t be punished by saying he asked AI not to hallucinate or make up the cases. Hard to believe.”

Another hinderance Champion has encountered in the legal world is “AI pollution,” where more and more content, he said, is being produced for the web and being scraped by AI. “It becomes this self-fulfilling cycle of BS, where it’s pulling the old BS to write new BS. The problem is people just blindly put it up because they are creating content.

“A lot of times, AI is not going into our legal data bases to pull cases, it’s looking on to the web. So, if you have false content on the internet, and you’re a lawyer or client searching legal advice that has not been properly vetted, you’re going to get bad advice.”

Though several state bar associations have released guidance on the use of AI for legal proceedings, the Idaho State Bar (ISB), according to Executive Director Maureen Braley, has yet to do so in the Gem State.

“We are working closely with the Idaho Supreme Court and other stakeholders on the issue, including through the Court’s recently established AI Governance Work Group,” she said.

Veteran attorney Brad Frazer, a partner with the firm of Hawley Troxell in Boise, said he’d welcome some form of AI guidance, should it originate with the state’s highest court.

“I would be more interested in that type of guidance if it came from the Idaho Supreme Court by way of amendment to the rules of civil procedure or the rules of attorney professional conduct,” he said.

On Nov. 1, 1986, the Idaho Rules of Professional Conduct (IRPC) became effective (with subsequent amendments) by order of the Idaho Supreme Court. The IRPC are based largely on the American Bar Association (ABA) Model Rules of Professional Conduct with some Idaho variations.

Additionally, Idaho Appellate Rule 11.2 specifically requires that an attorney ensure a filing is “well-grounded in fact and is warranted by existing law or a good faith argument for the extension, modification or reversal of existing law,” regardless of whether an attorney uses AI or not.

In 2024, the ABA reminded its roughly 400,000 members that ethics rules “require lawyers to vet and stand by their filings or risk being disciplined,” and that such obligations extend to “even an unintentional misstatement” produced through AI.

However, while not fully opposed to the use of AI in legal proceedings, Frazer does prefer a “judicious” approach to the technology.

“If good legal advice is adequate, then use an AI,” he said. “But remember, it’s all academic until you get sued.”

Frazer, who has practiced law since 1989, easily differentiates between what he sees as “adequate” and “exceptional” legal advice.

“If you need excellent legal advice, use a human being who can understand AI outputs sufficiently with respect to that area of law to prevent hallucinations from sneaking in,” he said. “But if you need world class, bet-the-farm legal advice, then use a human being who is a subject matter expert. At that point whether they use AI or not is almost irrelevant.”

Unfortunately, it’s becoming more common to see AI-assisted lawyering land more than its fair share of litigators in the court’s hotseat.

This past January, a federal judge presiding in Kohls v. Ellison, District of Minnesota, Jan. 10, 2025, barred a submission from an expert on the “dangers of AI and misinformation” that cited non-existent articles that had been fabricated by ChatGPT. To quote the court: “The irony.”

In Champion’s home state of Georgia, an appellate court earlier this summer vacated a trial court’s order that was found to have relied on at least two fictitious AI-generated cases.

“Owning a firm as I do, we start thinking about these tools and whether we formally utilize them as a firm or whether individuals are using them,” he said. “It makes me nervous because I don’t want to be that lawyer who becomes the subject of a news article or the sanction order because something was cited such as a case that doesn’t exist.”

Similarly, the use of AI by litigants who may not feel they need to use an attorney, or can’t afford one to represent them, provides another concern not lost on Champion.

“I think for some there is a danger of people saying, ‘I don’t need a lawyer, I can do that,’” he said. “But it probably will be more a problem in areas that are underserved, outside say personal injury, which is not underserved.”

For instance, a landlord and tenant dispute is an area in which people typically pay out-of-pocket for a lawyer.

“Someone might use AI to write a complaint or a response to a lawsuit,” Champion said. “And that might not be great. At the same time, representing yourself and not having a lawyer is not great either. How much worse off will they be? I don’t really have an answer.”

But the one answer he does champion in his firm is the absolute need for the human element to be present.

“I’ll take my local human intake person any day over using AI,” he said. “Again, I’m 43 years old but still think of myself as old school. Roll up your sleeves and talk to your client. They need connection, a human, empathetic voice. There is the old saying, ‘trust but verify.’ But I would say use cautiously and always verify.”