People v. Crabill
Colorado Office of the Presiding Disciplinary Judge · Colo. OPDJ · Colorado bar guidance
Conduct
Filed a ChatGPT motion with fabricated cites; discovered the fabrication before the hearing but said nothing, then blamed an intern.
Consequence
1 year and 1 day suspension (90 days active, remainder stayed); first US bar suspension for AI-hallucinated filings.
Lesson
The coverup was worse than the error. Disclose immediately on discovery; Rule 3.3 candor obligations are the only safe path.
Verified April 24, 2026
- Citation
- People v. Crabill, No. 23PDJ067 (Colo. OPDJ Nov. 22, 2023)
- Decided
- November 22, 2023
Summary
Attorney Zachariah Crabill used ChatGPT to generate case citations and submitted them in a motion without verification. He discovered the citations were fictitious before a hearing but did not alert the court or withdraw the motion. When confronted by the court, he falsely attributed the errors to a legal intern. Six days later, he filed an affidavit disclosing his ChatGPT use.
- AI tool:
- ChatGPT
- Sanction amount:
- 1 year + 1 day suspension; 90 days active
What sanction did the court impose?
Total suspension 1 year and 1 day; 90 days active suspension; remainder stayed on 2-year probation with conditions. Presiding Disciplinary Judge approved stipulation to discipline. Violations: Colo. RPCs 1.1 (competence) and 1.3 (diligence) for failing to verify; 3.3(a)(1) (candor to the tribunal) for failing to alert the court or withdraw the motion after discovering the fabrications; and 8.4(c) (dishonesty/misrepresentation) for falsely attributing the errors to an intern. The coverup was the aggravating factor that elevated the sanction from a reprimand to active suspension.
Why does People v. Crabill matter for law firms using AI?
People v. Crabill was the first active bar suspension in the United States for conduct arising from AI-hallucinated citations. The facts matter because they establish the operative lesson: the coverup was worse than the error. Filing fabricated citations violated Rules 1.1 and 1.3; the 90-day active suspension turned on the Rule 3.3 failure to correct and the Rule 8.4(c) misdirection to an intern. For attorneys evaluating an AI hallucination that has already reached a filing, Crabill is the cautionary case — prompt correction is the only defensible path.