As AI technology continues to advance, providing greater efficiency and the ability to process extensive amounts of data, legal professionals are increasingly tempted to utilize Generative AI as a convenient shortcut for drafting legal motions and pleadings. However, lawyers must always bear in mind the Rules of Professional Conduct and the State Bar Act when relying on Generative AI, or they may face serious consequences. To help inform legal practitioners of potential pitfalls, the Committee on Professional Responsibility and Conduct (COPRAC) of the State Bar of California recently published an opinion entitled, “Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law.”
In its opinion, COPRAC cautions that because each Generative AI model and product operates differently and incorporates data differently, there are risks to relying on any one model’s outputs for use in legal work. One area of particular concern is the lawyers’ duty to ensure client confidentiality. A lawyer shall not reveal client information protected from disclosure unless the client gives informed consent and must maintain inviolate the confidence, and at every peril to himself or herself to preserve the secrets, of his or her client. (Cal. Rules of Prof. Conduct, Rule 1.6.) A lawyer’s duty to preserve the confidentiality of client information involves public policies of paramount importance. (In Re Jordan (1974) 12 Cal.3d 575, 580.) Preserving the confidentiality of client information contributes to the trust that is the hallmark of the lawyer-client relationship.
Generative AI utilizes both user prompts and documentation uploaded to its system and integrates and shares third-party documentation and information in its outputs. Because this could result in inadvertent disclosure of client information, COPRAC recommends that lawyers consult with an IT expert regarding the security, confidentiality, and data retention protocols of an AI product prior to its use. Lawyers should never input any confidential client information into any generative AI solution without first ensuring there are adequate confidentiality and security protections. A lawyer should never use confidential client information with a generative AI product that uses inputted information to train its AI model or to provide responses to queries from other users. (See the COPRAC Opinion, Practical Guidance & Applicable Authorities.)
Another consideration when using a Generative AI system is that the outputs it generates could be false, biased, or incomplete, and reliance upon these outputs could result in lawyers violating the duty of competence. Competence is defined as applying the learning and skill, and mental, emotional, and physical ability reasonably necessary for the performance of a legal service, which includes the duty to keep abreast of the changes in the law and its practice, including the benefits and risks associated with relevant technology. (Rule 1.1, Comment [1].) Thus, COPRAC cautions that prior to using any Generative AI, a lawyer should understand to a reasonable degree how the technology works, including its limitations as well as the policies governing its use.
As with any tool, COPRAC recommends that AI-generated outputs should not be relied upon as a substitute for individual review and analysis. At best, these Generative AI products should only be used as a base from which a lawyer adds his or her own critical analysis to ensure accuracy, reduce bias, and ensure sufficient client protections.
For the complete COPRAC Opinion and Practical Guidance, see COPRAC’s website: https://www.calbar.ca.gov/Portals/0/documents/ethics/Generative-AI-Practical-Guidance.pdf