Donovan v. Clark County Assessor
Indiana Tax Court · Ind. Tax Ct. · Indiana bar guidance
Conduct
Pro se petitioner confirmed at oral argument she used generative AI to draft briefs; one cite was fabricated, three real cites misquoted.
Consequence
Petitioners admonished for failing to verify AI-generated citations. Tax assessment affirmed. No monetary sanction.
Lesson
Ind. Tax Ct. is the third Indiana appellate forum after Williams v. Kirch and Cingel to admonish pro se litigants for AI-fabricated cites.
Verified May 14, 2026
- Citation
- Donovan v. Clark Cnty. Assessor, No. 25T-TA-00002 (Ind. Tax Ct. Dec. 22, 2025) (McAdam, J.)
- Decided
- December 22, 2025
Summary
Indiana Tax Court decision authored by Judge Justin McAdam affirming a property-tax assessment in favor of Clark County. In a separate section titled "Warning regarding the use of artificial intelligence," the court found that pro se petitioner Linda Donovan cited a case that does not exist, "Rawles v. Monroe Cnty. Ass'r, 48T10-1705-TA-00014, slip op. at 5-6 (Ind. Tax Ct. May 13, 2019)," and quoted language from three other Indiana Tax Court cases that does not appear in any of those opinions. At oral argument, Ms. Donovan confirmed she had relied on artificial intelligence to draft her briefs, leading the court to conclude the nonexistent citation was an AI hallucination. Citing Williams v. Kirch as the controlling Indiana appellate authority on AI-citation conduct, the court admonished the Donovans for failing to confirm the accuracy of their authorities but imposed no further penalty.
- AI tool:
- Generative artificial intelligence (specific tool not identified); use confirmed at oral argument
What sanction did the court impose?
Indiana Board of Tax Review's final determination affirmed on the merits. Pro se petitioners admonished for citing a fictitious case and misquoting three real cases. No monetary sanction.
Why does Donovan v. Clark County Assessor matter for law firms using AI?
Donovan extends Indiana’s emerging AI-citation case-law to the Tax Court, making Judge McAdam the third sitting Indiana appellate judge (after Vaidik in Williams v. Kirch and Felix in Cingel v. Ferreri) to issue a published admonishment for fabricated cites generated by artificial intelligence. The opinion is methodologically careful: it identifies the fabricated case (“Rawles v. Monroe Cnty. Ass’r”), traces the misquoted language back to the actual Indiana Tax Court cases the petitioners purported to cite, and pins the AI use to Ms. Donovan’s own oral-argument concession rather than inferring it.
For a firm representing clients in Indiana property-tax appeals, the operational implication is that the Tax Court has now joined the Court of Appeals in actively scrutinizing AI-assisted briefing. Even a pro se litigant assisting themselves with consumer AI tools faces public admonishment in a forum where the bench is small and judges remember prior parties. A firm that helps clients prepare informal pro se filings or that reviews briefs from clients before the Tax Court should treat AI-generated citations as a known failure mode and verify each authority against the official Indiana Tax Court opinion archive before submission.
Implications for your firm
Operational steps a firm reading this case may wish to consider documenting. Strategic and rule-application calls belong to your firm's attorneys.
- Document a litigation-support workflow that prohibits filing AI-generated citations without independent verification against a primary case-law database, even for pro se assistance projects or low-stakes tax matters.
- Train staff supporting clients in property-tax appeals that the Indiana Tax Court is now actively detecting AI hallucinations and will admonish pro se litigants on the record; pro se does not insulate a client from a public admonishment that may follow them in subsequent matters.
- Review any AI-assisted research outputs for misquotation of real cases, not just fabricated cites; Donovan misquoted three Indiana Tax Court opinions verbatim, which is harder to catch than a fully fabricated citation.