It is undisputed that technology has transformed the practice of law. Many attorneys remember when cases were looked up in books, not online legal databases. When briefs and pleadings were handwritten instead of typed. Some courts still do not use electronic filing, instead requiring everything to be done through hard copy. The legal profession is seeing the latest technological shift happen now, as artificial intelligence – specifically generative artificial intelligence – is slowly integrated into the profession. Generative artificial intelligence (“Gen AI”) is a type of artificial intelligence which actually creates content. It is designed to produce output, such as text, images, or even legal work product.  

As Gen AI becomes increasingly common in the legal field, attorneys should be aware of potential ethical pitfalls with its use. Many attorneys have heard of recent cases where Gen AI use has gone awry. For example, in 2023 an attorney was sanctioned after submitting a brief to the Southern District of New York containing cases generated by ChatGPT (a Gen AI chatbot) that did not exist. The attorneys who submitted the brief were sanctioned both for filing the brief and for continuing to defend the fake citations. Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023).

More recently, in April of this year, a plaintiff representing himself at the Appellate Division attempted to argue his case using a ChatGPT-generated avatar as his counsel. The avatar appeared over video as a man standing in a collared shirt and a sweater, prompting one of the judges to ask the plaintiff if the man was his attorney. The plaintiff later explained that he had created the avatar to “present my arguments in the most efficient manner possible,” after stumbling over his words in previous legal proceedings. Man Employs A.I. Avatar in Legal Appeal, and Judge Isn’t Amused, available at https://www.nytimes.com/2025/04/04/nyregion/ai-lawyer-replica-new-york.html).

Although the profession may be a long way from embracing AI-generated counsel, Gen AI offers tangible tools for the legal field in the areas of drafting, research, and document collection and production. Both Lexis and Westlaw, two of the biggest legal databases relied on by attorneys, offer their own Gen AI product. With these tools, attorneys can type in specific, focused research questions and receive an AI generated memorandum with citations to relevant cases. Attorneys can upload briefs for Gen AI to review for relevant cases that may have been missed or check whether the cited cases are still good law. Attorneys can also upload an opposing party’s brief and ask AI to analyze the work to identify relevant authority that may counter arguments raised. 

With all the increasingly sophisticated technology comes risks attorneys must know. First, attorneys need to be cognizant of whether the Gen AI they use relies on customer data to train its models. For example, if a draft brief with client information is uploaded to a model such as ChatGPT, that sensitive and privileged information may not remain confidential. This could violate ethical obligations by attorneys to maintain attorney client privilege. Other Gen AI products specifically built for attorney use, such as Lexis and WestLaw’s products, generally have additional layers of protection to ensure that drafts and other work product can be uploaded and analyzed while remaining confidential. For example, Lexis’s Gen AI product describes having a “fully encrypted Lexis+ AI environment” that “ensures data privacy protections that far exceed what consumer-grade AI tools offer.” Before uploading any work product or documents with confidential information to any type of Gen AI, an attorney should check the policies of the Gen AI or check with their IT department to ensure they are not violating attorney-client privilege by doing so. 

In response to cases of attorneys misusing Gen AI, the American Bar Association issued Formal Opinion 512 last year. The opinion offers guidance on the use of Gen AI tools and cautioned attorneys that they must fully consider their applicable ethical obligations when using such tools. The New York City Bar Association also issued Formal Bar Opinion 2024-5, regarding ethical obligation of lawyers and law firms relating to the use of Gen AI in the practice of law. 

There are several key takeaways from the ABA and New York City Bar Associations’ opinions, which raise many of the same concerns for attorneys. First, attorneys must ensure that client information will remain confidential before using Gen AI products. As discussed above, many Gen AI models do not offer any confidentiality for the information they are given. Second, duties of competence and diligence also require awareness that Gen AI systems are fallible. Attorneys cannot blindly rely on Gen AI’s output without checking to ensure accuracy, regardless of whether Gen AI is utilized to draft a contract provision or a memorandum of law. Third, law firms and managerial attorneys should create clear policies regarding permissible use of Gen AI. This allows employees to fully understand the restrictions imposed on them by their employer regarding allowable use of Gen AI. Fourth, to the extent that use of Gen AI makes an attorney more efficient, the attorney cannot charge hourly fees for the time a task would have taken without use of Gen AI.

Generative artificial intelligence is a growing tool increasingly relied on by the legal profession. While its growth in the legal profession should be fostered and encouraged, it also needs to be approached with an understanding of how it can best be used in a client’s best interests without running afoul of ethical obligations. 

Ryan Lefkowitz is an associate with Adams Leclair LLP and can be reached at RLefkowitz@adamsleclair.law.

This article originally appeared in The Daily Record.