COPRAC Advisory Regarding Artificial Intelligence (AI) Hallucinations
Due to the increased usage of artificial intelligence (AI) in the legal profession, the Committee on Professional Responsibility and Conduct (COPRAC) continues to provide guidance on relevant ethical and practical considerations that arise from the use of these technologies.
Generative AI tools (such as ChatGPT and Perplexity) are computer applications that can create text, images, or other content in response to user prompts. In the legal context, they may be used for tasks such as brainstorming, research, drafting, or summarizing information.
While these tools can be helpful in streamlining some aspects of legal work, attorneys must use them in a manner consistent with their duty of competence (rule 1.1), diligence (rule 1.3), and responsibilities as managerial and supervisory lawyers (rule 5.1). Competent use of such technology requires understanding its limitations, including the risk of fake or “hallucinated” content, outdated or incomplete legal authorities, and the inadvertent disclosure of confidential client information through prompts.
Courts have sanctioned attorneys for submitting AI-generated filings containing false or fabricated authorities, and an attorney’s lack of awareness of the risk of “hallucinated” content does not relieve the attorney of responsibility for ensuring the accuracy and integrity of any work product submitted. Attorneys must independently verify any AI-assisted work product before relying on it in any context. Diligent representation requires that attorneys not delegate their professional judgment to AI, but instead review, edit, and take responsibility for the substance and timing of all filings, communications, and advice.
Attorneys with managerial or supervisory authority must also implement reasonable policies, training, and oversight to ensure that any use of generative AI by attorneys does not compromise client confidentiality or replace appropriate legal analysis, supervision, or quality control. Ultimately, licensees should evaluate these tools thoughtfully, balancing their potential benefits while understanding the potential pitfalls.
COPRAC is actively working on revisions and updates to its practical guidance regarding AI. The updated practical guidance will be presented at the May 14–15, 2026, Board of Trustees meeting for approval. In addition, proposed amendments to the Rules of Professional Conduct are currently out for public comment. Licensees and members of the public are encouraged to submit written comments on the proposed amendments.
