Overview
Generative Artificial Intelligence (GAI), particularly Large Language Models (LLMs), are reshaping the way legal work is performed. These tools can streamline drafting, research, discovery review, and more. With the emergence of this new technology, it is essential to recall the professional rules which govern attorneys and apply to GAI tools. LLMs, such as ChatGPT, introduce new contexts in which long-standing ethical rules already apply. Recent court decisions across the country highlight the importance of using AI responsibly and in accordance with the ethical rules, and underscore the need for careful human oversight.
Competence and Diligence in an AI Environment
The American Bar Association Model Rules 1.1 (Competence) and 1.3 (Diligence) require that lawyers provide competent representation, including thorough preparation, and that they act with reasonable diligence. Since LLMs can “hallucinate” cases and text that appear authoritative, but may be fabricated, lawyers cannot take on cases in areas of law where they lack competence by relying on GAI. Even if competent in an area of law, lawyers must diligently review any AI output to make sure it is accurate. Further, they must make sure that they meet deadlines and review their submissions in sufficient time so they will not need to rely on AI to complete an assignment.
Several widely discussed cases demonstrate the consequences of neglecting these competence and diligence responsibilities. For example, in 2023, lawyers dodged sanctions when they submitted a brief that contained made-up cases. Since that case, the penalties for misuse of AI and neglect of professional responsibilities have increased. In January 2024, an attorney was referred to the Second Circuit Court of Appeals Grievance Panel for punishment after submitting a brief that contained a fake case citation.1 In another case, in the Southern District of New York, a law firm was fined $5,000 for submitting non-existent judicial opinions with fake quotes and citations created by ChatGPT.2 Around the country, attorneys have been removed from cases,3 ordered to pay litigation sanctions in the tens of thousands of dollars,4 and have been publicly reprimanded and referred to state bars for mis-using generative AI tools.5
These cases illustrate a consistent lesson, namely, that AI tools cannot replace human judgment, legal training, or careful research and review. Courts increasingly expect attorneys to certify that any AI use has been reviewed for accuracy, and several judges and courts have implemented standing orders requiring such verification. A few have even banned the use of AI-generated content. As a result, increased punishments and fines for attorneys who continue to misuse GAI in their work can be expected in the future.
Communication and Transparency
Model Rule 1.4 requires attorneys to keep clients promptly and reasonably informed so that they can make decisions about representation. This means that an attorney may be under an obligation to inform the client about the use of AI to prepare legal documents, such as drafting patent applications, briefs, etc. According to the American Bar Association Standing Committee on Ethics and Professional Responsibility Formal Opinion 512, issued on July 29, 2024, lawyers need not disclose every use of AI, but transparency may be necessary if AI plays a significant role in document generation or analysis. Clear communication and transparency ensure that clients understand both the benefits and the limitations of the technology.
Fees
According to Model Rule 1.5(b) an attorney taking on the representation of a client is responsible for communicating to the client the basis or rate of the fee and expenses that the client will be charged. This indication of the fees and expenses is preferably indicated in writing, before or within a reasonable time after commencing the representation, except when the lawyer will charge a regularly represented client on the same basis or rate. Any changes in the basis or rate of the fee or expenses shall also be communicated to the client.
With the advent of AI in legal representation, its effect on the proper fee a client should be charged has arisen and is still being worked out. Such questions as: “Should the attorney be able to charge the client for purchase of AI as an expense? Should any savings from use of AI be passed onto the client? Can a client demand a lower price because of the attorney’s use of AI?” come up. One thing is fairly clear. If the client is billed by the hour and the attorney spends less time on the matter as a result of the use of AI tools, billing the hourly charge without adjustment is not acceptable.
Confidentiality of Information
According to Model Rule 1.6 “(a) A lawyer shall not reveal information relating to the representation of a client unless the client gives informed consent, the disclosure is impliedly authorized in order to carry out the representation …” Further, “(c) A lawyer shall make reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client.”
Use of LLM GenAI tools raises data protection and privacy issues. Some AI tools use the inputs or prompts of users to further train the tool. If the AI tool is “open” or publicly accessible, such use may disclose the client’s information. Thus, using a client’s confidential information in a prompt, e.g., to generate a patent application, may result in its disclosure. The same may be true of the use of client confidential information to train the AI. Thus, in evaluating AI tools, the attorney should determine whether input data is used to train the AI, and if so, use caution in the types of information that is used in prompts. Alternatively, the attorney can restrict the use of AI tools to those that do not train on input data, and even better, to those that are closed and only available for internal use.
Supervision and Law Firm Obligations
Model Rules 5.1 and 5.3 require partners and supervisory attorneys to ensure that lawyers and non-lawyer staff comply with ethical obligations. This includes firm-wide policies governing how and when AI may be used, and how confidential information is protected. Thoughtful supervision helps ensure that use of GAI tools contribute to efficiency, while maintaining professional responsibilities.
USPTO Considerations
USPTO Rules 11.101 and 11.103 require a practitioner registered to practice before the office, (which includes attorneys and non-attorney agents) to “provide competent representation to a client,” which “requires the legal, scientific, and technical knowledge, skill, thoroughness and preparation reasonably necessary for the representation.” These rules further require that practitioners “act with reasonable diligence and promptness in representing a client.” Thus, these rules are similar to Model Rules 1.1 and 1.3.
In the patent context, the USPTO has issued guidance emphasizing practitioners’ duty to determine proper inventorship in situations involving AI-assisted inventions. Examiners may require information about whether AI contributed to the conception of the claimed subject matter or other elements. As with other areas of practice, the duty of reasonable inquiry applies regardless of whether AI tools were used to draft or review submissions.
Conclusion
It is not sufficient to try to avoid the risks stemming from GAI by refusing to use it or learn about it. In fact, Comment 8 to Model Rule 1.1 requires that lawyers keep abreast of changes in the law and its practices “including the benefits and risks associated with relevant technology…” Thus, at a minimum, lawyers have a duty to understand how AI functions and how it can be used or mis-used.
AI has the potential to enhance legal services, including decreasing the cost and time, and potentially improving work product and workflow. Despite the benefits, there are risks to using AI, and the ethical rules obligate attorneys to understand such risk and benefits, and to use these tools responsibly. The cases emerging from around the country make clear that courts will hold attorneys accountable for inaccuracies and violations resulting from the misuse of AI tools. Ultimately, the ethical use of AI rests on the same core principles that have always guided the legal profession: competent and diligent representation, timely and thorough review and verification, transparent communication, and commitment to protecting client interests.
1 Park v. Kim, —F.4th—, No. 22-2057, 2024 WL 332478, at *2–4 (2d Cir. Jan. 30, 2024).
2 Mata v. Avianca Inc., —F. Supp. 3d —, No. 22-CV-1461, 2023 WL 4114965, at *1 (S.D.N.Y. June 22, 2023).
3 Magpul Industries Corp. v. Mission First Tactical, LLC, No. 24-5551-KSM (E.D. Pa. Apr. 28, 2025).
4 Lacey v. State Farm General Insurance Co., No. CV 24-5205 FMO (MAAx) (C.D. Cal. May 6, 2025).
5 Johnson v. Dunn, No. 2:21-cv-1701-AMM (N.D. Ala. July 2025).
Professionals
Melvin Garner
Partner
Audrey Trace
Associate
