Stevens v. BJC Health System
Missouri Court of Appeals, Eastern District, Division Three · Mo. Ct. App. E.D. · Missouri bar guidance
Verified May 5, 2026
- Citation
- Stevens v. BJC Health Sys., No. ED112759 (Mo. Ct. App. E.D. Mar. 18, 2025)
- Decided
- March 18, 2025
Summary
Appellant Angela Stevens, proceeding pro se, appealed from a decision of the Labor and Industrial Relations Commission affirming the dismissal of her unemployment-benefits appeal after she missed a telephone hearing. Her appellate brief contained citations to six cases that the court was unable to locate. The court found that the fictitious citations suggested the use of AI in drafting the brief, quoting Kruse v. Karlen and Mata v. Avianca for the proposition that fake opinions are 'not existing law' and that reliance on them is 'an abuse of the adversary system.'
- AI tool:
- AI (inferred by the court; not acknowledged by the appellant)
What sanction did the court impose?
The court affirmed the Commission's decision. No monetary sanction was imposed in this case. In a footnote, the panel issued a forward-looking warning 'in light of artificial intelligence's increasing prevalence': 'using artificial intelligence to draft a legal document may lead to sanctions if the user fails to perform a critical review of the end-product to ensure that fictitious legal authorities or citations do not appear in filings with this Court or any other court.' Opinion by Judge Renee D. Hardin-Tammons; Hess P.J. and Gaertner J. concurring.
Why does Stevens v. BJC Health System matter for law firms using AI?
Stevens v. BJC Health System follows Kruse v. Karlen by approximately one year and represents a refinement in how Missouri courts treat AI hallucinations in appellate practice. Where Kruse involved near-wholesale fabrication, Stevens involved six unlocatable citations and required the court to infer AI use. The opinion shifts the burden: it is no longer whether a party used AI, but whether the party performed adequate review, and failure to review is itself sanctionable going forward. The court’s characterization of fabricated citations as “an abuse of the adversary system” elevates AI hallucinations from a technical deficiency to an ethical violation.