Reprinted from Law.com
Built Different: Are Litigators the Legal Industry’s Biggest Hurdle in AI Uptake?
Litigators’ cautious embrace of AI reflects courtroom scrutiny, ethical risks and a focus on better work, not blanket resistance to the technology.
This article is part of our content offering surrounding Pathways: Future of the Business of Law 2040. For more information on Pathways, check out this article from group chief content officer Gina Passarella and see the editor’s note below.
Litigators are often cast as the legal industry’s holdouts on artificial intelligence. Lawyers and law firm technology leaders who work with them say the real obstacles are fear, slow-moving courts, and the risk of putting their names on every court filing.
“The real reason why there’s a lot of hesitancy to adopt AI for all lawyers, not just litigators, is fear,” said David Wang, chief innovation officer at Cooley in Palo Alto, California.
Sarah McCormick, director of legal service delivery in the practice innovation department at Honigman in Detroit, said litigators aren’t resisting AI so much as using it differently and more cautiously.
“Lawyers adopt tools they can stand behind,” she said. The stakes for litigators are unusually public.
In 2023, a federal judge in Manhattan fined two lawyers and their firm $5,000 in Mata v. Avianca for submitting a brief that cited fake cases invented by ChatGPT. As of Thursday, courts had flagged AI-fabricated material in 1,553 U.S. cases, according to a global database kept by lawyer and researcher Damien Charlotin. Many of those filings came from self-represented litigants, not lawyers. And some judges now require attorneys to disclose when they use AI.
At the same time, the American Bar Association has said a lawyer’s duty of competence includes understanding the technology’s benefits and risks.
A Structural Gap
Wang said part of the gap is structural. Corporate lawyers work closely with business clients that use AI every day and expect their lawyers to keep up. At Cooley, many of those clients are technology companies.
Litigators answer to the court system instead, where judges and clerks “are not exactly early adopters,” Wang said. Nothing pulls litigators toward AI the way clients pull deal lawyers, he said.
Frank Ramos, a partner at Goldberg Segalla in Miami, who does mostly civil defense work, sees some of that pull from his insurance and corporate clients. Most now expect their law firms to use AI, he said, but they aren’t paying for it. Firms are absorbing the cost and pitching their AI use as a selling point.
“I haven’t seen clients volunteer to pay for it,” Ramos said.
Every filing is also what Wang called “a crystallizing moment”: a document a lawyer signs and vouches for in court. Widely publicized cases of AI-invented citations have stuck in litigators’ minds, he said, and “lawyers don’t like risk.
‘Every Mistake Is Public’
Jonathan D. Bick, a Roseland, New Jersey, attorney at Brach Eichler, who teaches at Rutgers Law School, said litigators know opposing counsel, judges and clerks will scrutinize everything they file. “Any error that they make will not only be highly visible but could be costly,” Bick said.
Court rules add another layer. Bick said some courts and judges in New York, New Jersey and Pennsylvania have standing orders requiring lawyers to disclose AI use and certify they checked the work themselves. A lawyer who doesn’t track every judge’s latest order can avoid breaking one by skipping AI entirely, he said.
Ramos said a steady stream of sanctions orders has led many lawyers to the same conclusion. “There are several lawyers, not a small subset of them, who figure, ‘Well, if I don’t use AI, I can’t get in trouble for using AI,'” he said.
Wang argues the most famous fake-citation case was about the lawyer, not the tool. “That wasn’t an AI problem. It’s a human problem,” he said.
A lawyer who files work without reading it is making the same mistake whether the draft came from a junior associate or a chatbot, he said.
Caution, Not Resistance
McCormick at Honigman said litigators work “basically under a microscope” when it comes to using AI. “It’s not that the litigators aren’t using the AI tools. It’s just that they’re using them with a lot more caution or they’re using them in different ways than you would see the transactional lawyers use them,” she said.
She said litigators are more tech-forward than many other practice groups, a habit that dates to the early days of e-discovery and electronic document review. Much of the AI they use now arrives as add-ons to platforms they already know.
When Honigman first rolled out a general purpose AI tool to a small group, litigators’ demand for it helped make the business case for buying more licenses, McCormick said. Pilots with litigators don’t fail more often, she said, but they come with extra rules. Litigation files can carry “attorneys’ eyes only” designations, and court rules bar some documents from going into AI tools at all.
McCormick, who serves on the American Bar Association’s AI Legal Ethics Working Group, said that caution is justified. Litigators “have good reason to … demand that the AI tools are outputting correct information before it reaches a judge,” she said.
She also sees it as useful pressure on the companies that build the tools. “Their skepticism really helps everyone,” she said. “It helps the firm, it helps clients, it helps make the tools better because it forces the tools to be even more precise than they already are.”
Ramos said part of the lag is cultural. Lawyers and law firms “tend to be slow adopters of technology,” he said, and many trial lawyers worry AI will dull a skill they consider an art. “A lot of trial lawyers feel that it’s a very creative process. It’s a very almost artistic process, and it’s a craft,” he said.
‘I Reject in Its Entirety’
Matt Wetherington, a trial attorney in Atlanta, rejects the premise that litigators are holding AI back. “I reject in its entirety the idea that there are a large number of lawyers that are actively standing in the way of AI or rejecting the use of it,” he said. “I think your average good litigator is going to use literally every single tool and resource available to them.”
He said surveys on lawyers’ AI use can’t be trusted, because lawyers and firms may not answer honestly while ethical questions linger.
“If you were to ask litigators, ‘Are you using AI?’ I think that they may not tell the truth,” Wetherington said. “I don’t think you’re getting an honest poll at this point.”
The bigger lag, he said, is on the bench. Most of the judges he works with on judiciary committees don’t use AI daily. He said a recent poll of Georgia judges found most have no internal AI policy for their clerks and staff. A separate survey of federal judges, led by Northwestern University researchers with the New York City Bar Association, found that about one in four respondents had no official AI policy for their chambers.
He called rules requiring lawyers to disclose AI use “an exception that would eat the rule,” because AI is already built into the tools lawyers use every day. “Westlaw has AI, Lexis has AI, Microsoft Word has AI,” he said.
Wetherington uses AI heavily in his own practice, from deposition prep to real-time witness fact-checking. His firm built its own tools because of privacy concerns.
“AI tools make preparing for a deposition dramatically better—not necessarily easier, but better,” he said.
During one deposition prep, he said, the AI asked why he wasn’t questioning the witness about a particular document. He thought it had never been produced. It had been, as a file embedded inside a PDF he had missed.
On hallucinations, Wetherington said he hasn’t seen a legal hallucination from a frontier model in about a year and a half. But his rule hasn’t changed: “You still got to read the cases.
For all that, he is no AI booster. “I think within three years AI is going to ruin the legal industry absolutely, and I hate it,” he said.
A Duty to Use It?
Wang put the question in Silicon Valley terms. Asking whether a lawyer uses AI in 2026, he said, is “like saying, ‘Hey, do you use electricity or do you not use electricity?'” The real question is what a lawyer uses it for.
Bick goes further. He points to ethics rules that require lawyers to keep up with technology that could help them represent clients. “I think that it may violate professional standards and then actually be an ethical violation not to use AI,” Bick said.
Wang said lawyers who haven’t tried AI may already fall short of the technology competence requirement many states have adopted. “I don’t think you’re meeting that if, in 2026, you haven’t even tried,” he said.
Ramos said Florida’s ethics guidance on AI, like that of other states he has reviewed, assumes lawyers will understand the technology and use it where it helps clients. He said lawyers who refuse to use AI could face malpractice exposure within two to three years, once clients argue that skipping the technology cost them a better result. “I think it’s actually a closer window,” he said when asked whether that was five years away. For now, he said, the malpractice claims he has seen involve lawyers who misused AI, not ones who avoided it.
Using AI carries its own ethical duties, Bick said. He said lawyers should tell clients they use AI and get their permission, preferably in writing. He also warned that many general purpose AI tools feed users’ prompts back into their training, which could expose confidential client information.
Better, Not Faster
Wang called studies on AI time savings “kind of a red herring.” In high-stakes litigation, he said, deadlines are tight and what matters is the quality of the work. “Will there be efficiencies? Absolutely. But that’s incidental,” he said.
AI lets a legal team review more evidence, refine and rehearse arguments more times, and train junior lawyers better than it could otherwise, Wang said. That leads to better outcomes for clients.
“If you’re a litigator, surely everybody can agree that what you care about is winning,” he said. Lawyers who are as good with a pad and pencil and also use AI, he said, “eventually … will just win more.”
Wetherington said the same about his deposition work. “I don’t know that it saved me any time,” he said. “If anything, it’s taken more time.”
Ramos uses AI to prepare for trial, including slides, direct and cross-examinations, and deciding which jurors to strike. His firm uses Harvey, Microsoft Copilot and LexisNexis’ Protégé.
“I find that lawyers who use it ethically, safely and effectively are just better trial lawyers,” he said. He acknowledged the tension with the billable hour.
“There’s a paradox where you’re spending more money to earn less money,” Ramos said.
Firms that adopt AI early will bill fewer hours per case but attract more work, he said. He also expects more flat-fee billing by case or by stage, especially in predictable matters such as premises liability and auto cases.
Training Judgment, Not Clicks
Bick advises hesitant litigators to take a continuing legal education course on AI best practices, rethink how they bill for work AI now speeds up, and read the ethics guidance from the ABA and their state bar.
Wang said the lack of training is often an excuse. “This is literally the most easy-to-use technology in the history of technology,” he said. “There’s no buttons, there’s no menus.” Using it well is another matter, Wang said, and that takes significant training. At Cooley’s first-year associate training program, which the firm calls Cooley College, the AI segment came first.
Wang’s example for new associates: Don’t paste a partner’s confusing email into an AI tool and send back whatever it produces. “That’s the training. It’s not how to use the tool,” he said. “It’s using your judgment on when you should use the tool and when you shouldn’t use the tool.”
Taking AI away from junior lawyers out of fear they’ll misuse it only means “they will never learn how to use the tool appropriately,” Wang said.
McCormick said firms share a concern that younger associates who lean on AI too much “won’t learn how to do their actual jobs.” Honigman teaches them to use it while still learning the work, she said, because AI “is not going anywhere.”
Ramos expects AI to take over much of the work traditionally given to first- and second-year lawyers within the next 18 to 36 months. That includes basic motions, discovery requests, research memos and record summaries, which would leave junior lawyers checking the output instead of producing it. Young associates have to learn the craft the traditional way while learning the tools at the same time, he said, because they can’t yet judge what AI gives them. “Someone who’s just starting, they have no idea. I mean, it looks good to them,” he said.
Firms also need rules. Not having a firm-level AI policy in 2026 is “negligent, in my opinion,” Wang said.
Ramos, who speaks about AI a couple of times a week, said he is still surprised by how many lawyers in his audience are “at or near the starting block.”
Wang said fear is especially strong among senior lawyers, who are “very used to being experts” and don’t want to look like beginners. “It’s not too late. Start now, because you’re going to look even more foolish five years from now if you maintain that position,” he said. “If there’s one thing that litigators are known for, surely, it’s being fearless.”