Newsletters
Two U.S. federal judges acknowledged that staff members used artificial intelligence to assist in drafting recent court orders, which were criticized by Senate Judiciary Committee Chairman Chuck Grassley as “error-ridden.”
U.S. District Judge Henry Wingate of Mississippi and U.S. District Judge Julien Xavier, Neals of New Jersey revealed in letters released Thursday that the decisions bypassed standard review processes. Both judges have since implemented measures to improve oversight and ensure accuracy.
Neals noted that a draft decision in a securities lawsuit “was released in error–human error–and withdrawn immediately.” A law school intern had used OpenAI’s ChatGPT without authorization. Neals’ chamber has now created a formal AI policy and strengthened review procedures.
Wingate said a law clerk used Perplexity AI to synthesize publicly available information as a drafting assistant, and posting the draft was “a lapse in human oversight.” The original civil rights order was removed and replaced.
Grassley praised the judges for acknowledging the mistakes and urged the judiciary to adopt stricter AI guidelines, emphasizing that generative AI must not compromise litigants’ rights or fair treatment under the law.
Lawyers nationwide have also faced scrutiny for misusing AI, with fines or sanctions imposed in dozens of cases when AI output was not properly vetted.
3 min read
