In March of this year, a Missouri court partially hinged its ruling on counsel’s misuse of generative artificial intelligence (“AI”).
In Stevens v. BJC Health Sys., 710 S.W.3d 602, 604 (Mo. Ct. App. 2025), the Court, in its only Footnote, stated, “Appellant’s brief contains citations to six cases that this Court was unable to locate” and that citations to non-existent legal authority suggest the use of AI. Id. at 605. The Court then reminded the parties that the use of fake citations in persuading a court or opposing an adversary is an “abuse of the adversary system.” Id. However, arguably the most important statement from the Court regarding AI and professional liability is its’ warning that the use of AI in drafting legal documents may lead to sanctions if the party fails to perform a “critical review of the end-product to ensure non-existing legal authorities do not appear in court filings.” Id.
Reading between the lines, this case provides useful insights into AI and professional liability for Missouri legal practitioners. First, it appears that the Court views citations to non-existent legal authority as a presumption of the use of AI. Second, the Court considers citations to non-existent legal authority as an abuse of the adversary system not only in court filings, but in any attempt to “persuade the court or adversary.” Therefore, it follows that the Court takes issue with citations to non-existent legal authority in any persuasive correspondence to opposing counsel, not only in correspondence to the Court.
So, what is the bright line rule from this case? Simple. If the document or communication is persuading an adversary, persuading the court, or being filed with the court, ensure the legal authority in the citations exists.
