Guardrails Needed on AI Usage in the Legal Profession

By Jillian LeMaster-Dwyer, Esq.

Artificial intelligence is rapidly changing the practice of law. From legal research platforms to generative AI tools capable of drafting motions, summarizing discovery, and answering legal questions in seconds, the technology offers undeniable benefits. For solo and small firm practitioners in particular, AI presents opportunities to improve efficiency, reduce costs, and potentially expand access to justice for individuals who otherwise could not afford legal representation. 

However, alongside those benefits comes a growing concern: are AI systems operating as de facto attorneys without regulation, accountability, or oversight? 

The legal profession is built upon ethical duties, professional licensing standards, and accountability to clients and courts. AI systems, however sophisticated, are not subject to those same obligations. Yet many consumers increasingly rely on AI platforms for legal guidance, often without understanding the risks involved. 

One significant concern is accuracy, or the lack thereof. AI systems are known to generate false or misleading information with alarming confidence. Courts across the country have already encountered filings citing entirely fabricated cases generated by AI. In some instances, attorneys have faced sanctions after submitting briefs containing non-existent legal authorities produced by generative AI tools. 

Even more troubling are situations in which AI-generated misinformation may be influencing legal outcomes. In the case In re Domestic Partnership of Campos & Munoz, the law was altered because of on an AI “hallucination.” In this case, an attorney cited multiple fabricated cases, and the Court wound up including some of those AI-generated cases in its decision; essentially making those fabricated cases a reality.  

Those affected, in a circumstance like that in Munoz, have the ability to obtain compensation from the attorney who may have violated ethical standards. The same thing cannot be said for those who may use AI in lieu of legal representation. While AI can mimic persuasive legal writing, it lacks the capacity to verify truth, understand context, or exercise professional judgment in the way licensed attorneys are ethically required to do. 

Another emerging issue involves attorney-client privilege. Many clients assume that communications with AI systems are private and confidential, just as attorney-client communications are. They are not. 

In United States v. Heppner, a client used the AI platform Claude to analyze his legal situation, outline potential defense strategies, and develop legal arguments. The court determined that those communications were not privileged. This creates a substantial trap for consumers who may unknowingly waive confidentiality protections while attempting to obtain inexpensive or immediate legal guidance through AI platforms. 

The implications for practicing attorneys are significant. Clients may independently upload confidential documents, attorney-client communications, draft motions, or litigation strategies into AI programs without understanding that doing so could jeopardize privilege protections. For example, if an attorney provides a draft motion to a client and the client uploads that draft into an AI system for “suggested edits,” several problems immediately arise. First, the confidentiality of the draft may be compromised. Second, the AI-generated revisions may increase litigation costs because counsel must then review and analyze a potentially extensive list of AI-generated suggestions, many of which may be legally unsound or strategically counterproductive. 

AI usage may also complicate the attorney-client relationship itself. Attorneys are already seeing situations in which clients present pages of AI-generated legal theories, questions, or arguments for review. While client engagement is generally positive, excessive reliance on AI can substantially increase attorney time and fees. More concerning, clients may begin to distrust legal advice when it conflicts with the recommendations generated by a chatbot trained on generalized internet data rather than case-specific, or state-specific, legal analysis. 

The lack of accountability is perhaps the most troubling aspect of all. If a non-attorney human provides legal advice without a license, that individual may face civil or criminal penalties for the unauthorized practice of law. AI systems face no such consequences. They cannot be sanctioned, disciplined, suspended, or held ethically accountable. Yet they increasingly occupy spaces traditionally reserved for licensed professionals. 

None of this means AI should be abandoned. Used responsibly, AI can serve as a valuable tool for attorneys by improving efficiency and expanding access to information. But the profession must recognize that AI is not a substitute for legal judgment, ethical obligations, or the protections afforded by the attorney-client relationship. 

As AI becomes more integrated into everyday legal practice, meaningful guardrails are essential. Attorneys must educate clients about the risks of uploading confidential information into AI platforms. Courts and bar associations must continue developing ethical guidance governing AI usage. Lawmakers may ultimately need to confront the increasingly blurred line between technological assistance and the unauthorized practice of law. 

The legal profession has always evolved alongside technology. The challenge now is ensuring that innovation does not outpace the ethical and professional safeguards designed to protect the public. 

At TLD Law, our AI Committee remains at the forefront of evolving technology to help ensure our attorneys and staff are equipped with the training, policies, and tools necessary to responsibly leverage AI while protecting our clients from the risks associated with AI misuse. For questions about AI use in the legal field, please reach out to TLD Law at info@tldlaw.com.