NCSB Journal Summer 2026

Guarding Against AI Errors: Ethical Risks for NC Attorneys

B Y F R E D W. D E V O R E I I I A N D R O B W I L D E R

B y now, almost everyone has heard of AI, 1 “Artificial Intelligence.” 2 Although some folks refer to it as “almost intelligent” or “artificial ignorance.” AI has moved from theoretical curiosity to an everyday professional tool with aston ishing speed. In the legal profession, AI

now assists with research, drafting, document review, and case evaluation. Used prop

erly, it can enhance efficiency, reduce costs, and improve access to information.

However, when used improperly, AI can undermine professional judgment, violate

ethical duties, mislead courts, and expose attorneys to sanctions and malpractice lia

the-lightwriter/istockphoto.com

bility. When your name is on the signature line, you own every word.

with controlled databases and verifi cation safeguards. Many courts now require you to certify that, in con ducting your research, you either performed the AI searches in the legal databases or that you verified quota tions and citations obtained through AI searches outside of the legal databases. Discussions of “AI in law practice” often treat all uses of artificial intelligence as inter changeable, but that assumption is a mis w r

For those who are new to artificial intelli gence, there are several free websites that will provide “intelligent responses” to queries you propound. Many lawyers use publicly avail able AI platforms such as ChatGPT, Copilot, Gemini, or similar sites. Subscription versions of these sites offer expanded functionality. Meanwhile, major legal research providers—including Westlaw, Lexis, and Bloomberg—have incorporated AI tools into their platforms

take. In practice, lawyers encounter at least three distinct forms of AI. The first is public or general-purpose generative AI, such as ChatGPT, Gemini, or similar tools, which draw from broad, non-curated data sources and present the greatest risk of hallucinated

SUMMER 2026

8