Two attorneys in the New Orleans City Attorney’s Office have resigned after a federal judge discovered that a January motion they filed leaned on made-up case law generated by artificial intelligence. The filing – a Rule 12(b)(6) memorandum packed with nine nonexistent citations – triggered a tense show-cause hearing, monetary sanctions, and a scramble inside the office to rewrite its rules on AI. The incident has quickly turned into a local test case for how government lawyers can tap new research tools without blowing past their duty to verify what they put in front of a judge.
U.S. District Judge Carl Barbier issued an order sanctioning Assistant City Attorney Jalen Harris $250 and Deputy City Attorney James Roquemore $1,000, with an April 22 deadline to pay the clerk’s office, according to Justia. The order also required the lawyers to submit substitute pleadings and fix certificates of service the court found were missing.
How AI Turned A Brief Into Fiction
At a March 19 show-cause hearing, Assistant City Attorney Jalen Harris admitted that he started his research on Westlaw but then shifted to ChatGPT to speed things up, and never actually read or verified the nine cases the tool produced, according to WDSU. Harris, who joined the city attorney’s office in 2024, apologized repeatedly in court and promised not to make the same mistake again.
Judge Barbier’s order labeled the bogus authorities as “hallucinated” and noted that supervisory failures increased Roquemore’s share of the blame, pointing to his roughly 30 years of experience, according to Justia. The court declined to fine Chief Deputy City Attorney Corwin St. Raymond, but it did issue a formal admonition directed at the department.
Office Response And New Rules
City Attorney Charline Gipson, sworn in just days before the flawed motion was filed and whose name appeared on the signature block, told the court that her office convened a meeting and adopted a written policy on generative AI that took effect March 27. The policy requires disclosure of any prior AI use in work product and mandates annual certification of compliance, according to WDSU. The office said two lawyers submitted resignations in the wake of the episode, and staff members attended outside training on how to use AI properly…