CBA Record November-December 2025

THE YOUNG LAWYERS SECTION

AI’s Fast Lane and Avoiding the Roadblocks Ahead By Ion C. Moraru

A I is emerging in almost every corner of society at break neck speed. With the ability to peruse the entire Library of Congress or crank through massive datasets—all in the time it takes to make your morning tea—AI technologies are making inroads everywhere. In the intellectual property space, AI tools can swiftly perform prior art searches and analyses, assist with document drafting, detect and predict infringement, and provide asset valuations and market intelligence. Yet speed does not guarantee reliability or ethical safety. AI’s growth shows no signs of slowdown, and at the same time lawyers face increasing pressures to reach higher produc tivity. New companies are bringing state-of-the-art law-related AI products to the rescue, with the promise of revolutionizing the legal profession through streamlined tasks, reduced costs, enhanced accuracy, and strategic advantage. Despite little, if any, understanding of how these technologies actually work, many practitioners already incorporate AI into their legal work streams. Pitfalls and Sanctions But users beware! Many observers across legal circles are increasingly

concerned about the dangers of these yet-to-be-vetted technolo gies. For example, AI can generate work that looks impressively similar to the “real thing” or the “right thing,” but it utterly lacks any basis in reality or accuracy. Innocently designated, hallucina tions create false legal precedents, invent citations to nonexistent sources, and make seemingly credible arguments that lack ratio nal foundations. Federal courts have doled out plenty of sanctions to lawyers who fall short of Rule 11’s responsibilities (e.g., “thou shalt check your work before signing”). Even the courts themselves are not immune to the allure of this promising technology, but at least in court, an opposing party may set the record straight. And what of AI machinations that don’t undergo scrutiny? For instance, some tools store user information in unapproved locations and use it to train their models without permission, which can breach confidentiality and endanger attorney-client privilege. In the patent context, using an AI tool hosted on a for eign cloud-based server to draft content for a patent application may risk a host of penalties if the invention is “disclosed” with out a foreign filing license. Further, impacts on attorney certifica tion in a motion for summary judgment and what constitutes a

28 November/December 2025

Made with FlippingBook. PDF to flipbook with ease