Technology’s Effect on Advocacy
My Client’s Other Lawyer Is ChatGPT: What Attorneys Need to Know When Clients Turn to AI By Jade M. Brewster
With the rise of AI, clients now have unprecedented access to legal information that once required a law degree, or at least a Westlaw password. Increasingly, by the time a client reaches our office, they have already consulted their first legal advisor: artificial intelligence. They have asked whether they have a case, which causes of action to bring, what the defense will argue, what evidence they need, how much their case is worth, and what settlement they should accept. In some cases, AI has even supplied the relevant case law. Whether that case law actually exists is, of course, another question. Sometimes the answers are surprisingly good. Sometimes they are completely incorrect. And perhaps most dangerously, sometimes the advice is presented in a way that sounds extremely convincing. This presents an unusual challenge for lawyers. We must remember that legal information is not legal judgment. As AI becomes part of our clients’ everyday lives, lawyers need to understand not only how we use AI in our own practices, but how our clients are using it outside of personal and professional use. Because whether we like it or not, many of our clients have already consulted AI before they consult us.
AI does not necessarily know nuances about legal concepts that can make or break a case. Any experienced trial lawyer knows that case evaluation rarely turns on whether we can identify a theoretically viable cause of action. The complicated part is evaluating evidence, witnesses, credibility, damages, causation, defenses, venue, procedural obstacles, collectability, and the thousands other variables determining what happens to an actual lawsuit involving actual human beings. AI is remarkably good at providing an answer. Lawyers are supposed to be good at determining whether that answer survives contact with reality.
Hallucinations
Already, the legal profession learned a painful lesson in one of generative AI’s most famous shortcomings: hallucinations. AI can generate nonexistent cases, fictional quotations, inaccurate holdings, and citations that look legitimate enough to survive a quick glance. Courts across the country have dealt with lawyers submitting AI-generated authorities that did not exist. Noland v. Land of the Free, L.P., 2025 WL 2629868 (Cal. Ct. App. 2025); Malkeet Lnu v. Blanche, 177 F.4th 1014, 1024 (9th Cir. 2026). California attorneys have now been expressly warned about this problem. The State Bar of California’s Practical Guidance for the Use of Generative Artificial Intelligence recognizes that AI outputs may contain false, inaccurate, or biased information and instructs lawyers to critically scrutinize AI-generated work rather than rely on it as
Jade M. Brewster is a founding partner of Midtown Law Group LLP, located in Los Angeles, CA. Representing clients in employment and personal injury matters, Jade has been recognized by Best Lawyers and Super Lawyers.
jbrewster@midtownlawgrp.com
38
Consumer Attorneys of California
FORUM September/October 2026
Made with FlippingBook - Online catalogs