Generative AI, Plagiarism, and Copyright Infringement in Legal Documents

Abstract

Lawyers are increasingly using generative AI in their legal practice, especially for drafting motions and other documents they file with courts. As they use this new technology, many questions arise, especially surrounding lawyers’ ethical duties with respect to the use of generative AI. Headlines have been dominated by lawyers who have been disciplined for failure to confirm the output of generative AI systems, wherein the system hallucinates fake cases that the lawyers submit to opposing counsel and the court. Although that problem is certainly noteworthy, there are other potential issues for lawyers that have been more overlooked. The focus of this article is on two intriguing intellectual property questions that emerge when lawyers choose to use large language models like ChatGPT. First, might these lawyers be engaging in actionable, discipline-worthy plagiarism? This is unlikely to be the case, for several reasons, chief among them that copying and using boilerplate forms is standard practice in law. Nonetheless, courts and disciplinary agencies have reached surprisingly different conclusions on what counts as plagiarism in the practice of law and whether it is permissible. Any lawyer using generative AI should bear this in mind. Second, could these lawyers potentially be liable for copyright violations? Although this outcome may be unlikely, it is absolutely possible especially if lawyers do not understand that these tools can reproduce copyrighted text verbatim or if c…