Jason M. Hatfield, P.A. v. Pirani
U.S. District Court, Western District of Arkansas · W.D. Ark. · Arkansas bar guidance
Conduct
Used ChatGPT to draft post-trial motions citing nonexistent cases and quotations; signed and filed without verification.
Consequence
$1.67M in fee-shifting plus referral to Arkansas Office of Professional Conduct. Largest known US AI-hallucination sanction.
Lesson
Post-trial motions amplify exposure under fee-shifting. Courts use deterrence sanctions when they want to signal across a district.
Verified May 5, 2026
- Citation
- Jason M. Hatfield, P.A. v. Pirani, No. 5:22-CV-5110 (W.D. Ark.)
- Decided
- December 4, 2025
Summary
Attorney Tony Pirani (Pirani Law, Fayetteville) used ChatGPT to draft post-trial motions. The filings contained citations to nonexistent cases and quotations from nonexistent passages. Pirani electronically signed and filed the documents. He admitted the AI use at a pretrial conference on 2025-07-11; Judge Timothy L. Brooks issued an order to show cause on 2025-07-31.
- AI tool:
- ChatGPT
- Sanction amount:
- $1,578,172 attorney fees + $93,388 costs
What sanction did the court impose?
$1,578,172 in additional attorney fees and $93,388 in costs awarded to opposing party. Pirani was also referred to the Arkansas Judiciary's Office of Professional Conduct. Judge Brooks stated the sanctions were intended not only to address Pirani's conduct but also 'to deter similar misconduct by other attorneys practicing in the Western District of Arkansas, particularly with respect to the uncritical use of artificial intelligence in court filings.' The largest known dollar-amount AI-hallucination sanction in the United States to date.
Why does Jason M. Hatfield, P.A. v. Pirani matter for law firms using AI?
Hatfield v. Pirani is the largest AI-hallucination sanction in the country by dollar amount. The number itself is a signal: when a court wants to deter a category of conduct, it does not calibrate the sanction to the individual. Judge Brooks said so explicitly. The case also illustrates the downstream risk of unverified AI drafting — the sanctioned conduct was in post-trial motions, where a fee-shifting posture amplified the exposure.