With new AI tools popping up every day, attorneys can no longer trust that opposing counsel’s citations and propositions of law are accurate, and courts are not happy about it. Since May of 2023, courts in the United States have issued 212 sanctions decisions regarding generative AI-produced hallucinations in legal filings. 169 of those decisions were rendered this year. While AI pleadings are commonly used by pro se litigants, nearly half of these infractions were committed by legal professionals, including attorneys, paralegals, and judges.

Although some courts have been lenient, the use of AI without conducting a thorough review of all citations can expose a lawyer to significant risks. For example, a recent order from the District of Arizona sanctioned an attorney for the suspected use of AI in drafting the plaintiff’s opening brief. The attorney in question is the owner of a national social security disability practice and appears pro hoc vice in courts across the country. In this case, the sanctioned counsel received a brief from a contract attorney, which she apparently signed and filed without checking any of the citations or propositions of law.