3 www.loubar.org September 2026 PRESIDENT’S PAGE “ AI and the Practice of Law A brief that once took hours to draft, a lengthy cease-and-desist letter or hundreds of discovery requests can now be generated in minutes. The resulting work may look perfectly plausible to a client, a court or opposing counsel. But someone still has to read it. Did you know that Kentucky lawyers have an ethical duty to “keep abreast of the use of AI in the practice of law”? I didn’t, until I started this article. But that is exactly what KBA Ethics Opinion E-457 says. Kentucky attorneys must educate themselves about AI and understand its benefits and risks as part of their duty of competence under SCR 3.130(1.1). AI is quickly becoming an essential part of that technological competence. KBA E-457 was issued on March 15, 2024—a very long time in terms of AI development. In just two years, AI has gone from a novelty to something that can produce a reasonably convincing summary judgment motion in a few minutes. A few years ago, I tried a popular AI model and it was just ok. Today, it produces work on par with a bright and hardwork- ing, if perhaps overly eager, first-year associate—and it does so in minutes rather than hours. There is good reason for lawyers to embrace this technology. As noted in KBA E-457, AI can review and summarize thousands of pages of documents far faster than I can. It can identify patterns in large datasets, organize information, generate initial drafts and even help review and cite-check a final product. Used properly, AI can allow lawyers to spend less time on repetitive work and more time on the parts of practicing law that require judgment, strategy, creativity and human interaction. That should save money for clients and create new opportunities for lawyers. But it’s not all roses. AI can cite a case that does not exist, attribute a quotation to the wrong case or confidently state something that is simply incorrect. In its effort to prepare the argument it thinks you want, AI will take significant liber- ties with the facts and the law. Basically, AI is what happens when you take a bright and hardworking junior associate and tell him it is okay to fabricate facts and law as long as they seem consistent with the truth and suit your argument. E-457 emphasizes that lawyers using AI must verify its accuracy and remain responsible for their work product. The courts are willing to enforce that obligation. In May, the Kentucky Court of Appeals affirmed a family court order hold- ing a divorce litigant in criminal contempt for using AI to falsify medical records. The Sixth Circuit recently held, in a published decision, that attorneys practicing in the Circuit have an ethical duty to confirm the accuracy of their citations. Confidentiality presents another significant risk. Some AI tools may retain or use information entered into them, including confidential documents and privileged information. E-457 therefore advises lawyers to understand how an AI provider handles information before putting confidential client material into the system. The courts, again, have reached the same conclusion. Earlier this year, U.S. District Judge Jed Rakoff held that a criminal defendant’s legal communications with an AI program were not privileged. On July 31, 2026, the Supreme Court of Connecticut warned that attorneys risk breaching client confidences by sharing privileged or work product information with consumer AI programs. But the bigger issue—and one many of us are beginning to see—is the explosion of words in legal practice. Our profes- sion has always been a wordy one, but AI makes it remarkably easy to become an amateur Samuel Johnson. A brief that once took hours to draft, a lengthy cease-and-desist letter or hundreds of discovery requests can now be generated in minutes. The resulting work may look perfectly plausible to a client, a court or opposing counsel. But someone still has to read it. AI-generated mountains of legal prose create burdens for opposing counsel, clients and especially judges. If a lawyer uses AI to produce a 25-page brief, the lawyer still has an obligation to understand every argument, verify every authority and decide whether all 25 pages actually need to be there. More words do not necessarily mean better advocacy. Sometimes they just mean more work for everyone else. KBA E-457 addresses this through the lawyer’s duties of competence, supervision and review. But this may not be enough. The Campbell Family Court, for example, has adopted a rule requiring disclosure of whether AI was used to draft a court submission. Failure to comply “will result in a show cause hearing,” and potentially, referral for bar sanctions. The Campbell Family Court rule may well be adopted everywhere in short order. Some federal judges have begun en- tering standing orders requiring a certification on every filing of whether AI was used in that filing (see, for example, the April 2, 2026 Standing Order entered by U.S. District Judge Wendy W. Burger, M.D. Fla.). More are sure to follow. I would not be surprised if, within the next few years, we are all filing briefs and sending letters with a certification at the end stating the extent of AI use in creating the document. Such disclosures are a step in the right direction. In fact, I’ll go first. To help save time on this article, I asked a consumer AI program to summarize KBA E-457 for me. I’m not sure it saved much time—I ended up reading the opinion any- way—but it did a pretty good job. Samuel W. Wardle LBA President
