People v. Crabill
The case
People v. Crabill, Colorado Supreme Court, Presiding Disciplinary Judge, filed November 21, 2023 (approval effective November 22, 2023 per PDJ records), Case No. 23PDJ067. Copy of primary document: `~/workspace/goals/ai-sanctions-case-monitoring/files/sources/D-58-crabill-colorado-discipline.pdf` (the official filed Stipulation to Discipline, filed under the court's "FILED, Supreme Court, State of Colorado" stamp; the Colorado Supreme Court PDJ decisions page for 23PDJ067 lists it as the November 22, 2023 disposition).
What happened
Zachariah C. Crabill, licensed in Colorado in 2021, used ChatGPT to "bolster" the legal authorities in a motion to set aside a $31,350 default judgment. He did not read the cases ChatGPT supplied, and did not key cite, Shepardize, or otherwise verify their accuracy. The morning of the hearing he realized "all of my case cites from ChatGPT are garbage ... I can't even find the cases in Lexis," but he did not disclose the problem to the court. When the presiding judge raised concerns at the hearing, Crabill falsely blamed an intern for the bad citations rather than admitting he had used ChatGPT.
The rule and the sanction
Violations of Colorado Rules of Professional Conduct 1.1 (competence), 1.3 (diligence), 3.3(a)(1) (false statement to a tribunal), and 8.4(c) (conduct involving dishonesty), by stipulation under C.R.C.P. 242.19. Sanction: suspension for one year and one day, with all but 90 days stayed upon successful completion of a two-year probationary period. The stipulation also imposes costs of $224. Source URLs: the Law360 attachment copy of the filed stipulation served as the primary document (https://www.law360.com/pulse/articles/1770085/attachments/1); the Colorado Supreme Court PDJ decisions page lists the Crabill stipulation for 23PDJ067 dated November 22, 2023.
How it could have been avoided
Read and verify every citation an AI tool suggests before filing, using an authoritative database. Disclose errors to the court immediately upon discovery instead of deflecting blame onto others. Ask a colleague or supervisor for help when an assignment is outside your experience rather than leaning on an unverified tool.