⚡ CASE DIGEST
People v. Crabill — Colorado Presiding Disciplinary Judge, November 22, 2023
Colorado attorney Zachariah C. Crabill filed a ChatGPT-generated motion containing fictitious cases — then told the trial judge the fake research came from an intern, not AI. The Colorado Presiding Disciplinary Judge imposed a 90-day actual suspension plus a stayed suspension with two years’ probation — one of the first formal attorney discipline decisions for AI hallucination in the United States.
Category: Hallucinations & Sanctions | Jurisdiction: USA (Colorado) | Read time: 6 min
① Case at a Glance
| Full Citation | People v. Crabill, 23PDJ067 (Colo. PDJ Nov. 22, 2023) |
| Court | Colorado Presiding Disciplinary Judge (Supreme Court) |
| Date | November 22, 2023 |
| Category | Hallucinations & Sanctions / Attorney Discipline |
| Jurisdiction | USA (Colorado) |
| AI Tool Used | ChatGPT |
| Source | Colorado Legal Regulation Office |
② Background
In 2023, Colorado attorney Zachariah C. Crabill filed a motion to set aside a judgment in which he cited multiple case authorities generated by ChatGPT. Those authorities did not exist. When the trial court flagged the non-existent citations, Crabill told the judge the research had been prepared by an intern — a false statement designed to deflect responsibility. The deception was discovered. Crabill ultimately admitted he had used ChatGPT, failed to verify the output, and then lied to the court when confronted. The matter was referred to the Colorado Attorney Regulation Counsel, which filed disciplinary proceedings.
③ The AI Issue
The disciplinary proceeding raised two overlapping questions: (1) whether using unverified AI-generated citations violates the Colorado Rules of Professional Conduct; and (2) whether the subsequent lie — blaming an intern — constituted a separate, more serious violation of the duty of candour to the tribunal.
④ What the Court Decided
- Crabill violated Colo. RPC 1.1 (competence — failing to understand and properly use AI tools).
- He violated Colo. RPC 3.3(a)(1) (candour to tribunal — citing non-existent cases and lying about the intern).
- He violated Colo. RPC 8.4(c) (dishonesty, fraud, deceit, or misrepresentation) by fabricating the intern explanation.
- Sanction: 90-day actual suspension + 276-day stayed suspension + 2 years of probation with supervised filings and mandatory CLE on AI in legal practice.
⑤ Key Quote
“Respondent’s decision to blame a fictitious intern for the ChatGPT-generated research compounded the original misconduct with a second, deliberate act of dishonesty toward the court.”
— Colorado Presiding Disciplinary Judge, People v. Crabill, 23PDJ067
⑥ The India Angle
The Crabill case is a masterclass in how one error compounds into multiple disciplinary violations. Under the Advocates Act, 1961, BCI Rule 22 prohibits an advocate from “knowingly making a false statement of fact or law to the court.” The intern-blame tactic in India would constitute a false statement to a judicial officer — potentially actionable under Section 182 BNS (giving false information to a public servant). A BCI Disciplinary Committee under Section 35 can impose the same range of penalties: reprimand, suspension, or removal from the roll. An advocate caught making false statements to courts can also face action under the Contempt of Courts Act, 1971, on top of Bar Council proceedings. If you discover after filing that a citation is incorrect — the only defensible path is immediate disclosure. Crabill’s case shows that the cover-up is invariably worse than the crime.
⑦ Quick Takeaways
- AI hallucinations alone can trigger suspension — lying about them triggers a far harsher one.
- Blaming subordinates or fabricating explanations is treated as a separate, aggravated violation.
- Probation with supervised filings means every document filed is reviewed — an enormous practical burden.