BIRMINGHAM, Ala. — Three attorneys representing Alabama’s prison system faced sanctions and were removed from a federal lawsuit after a judge determined they had included fictitious legal citations in their court documents, generated by the AI tool ChatGPT.
U.S. District Judge Anna Manasco issued a stern order, reprimanding attorneys Matthew Reeves, William Cranford, and William Lunsford from the Butler Snow law firm for referencing nonexistent court cases. The judge characterized these citations as “hallucinations” produced by the artificial intelligence software, stating plainly that the references were entirely fabricated.
The lawsuit in question was initiated by an inmate at the Donaldson Correctional Facility, who claimed to have been repeatedly assaulted due to inadequate safety measures in the prison. After opposing counsel raised concerns over suspicious citations in May, Judge Manasco sought to verify the cases and found no records of them. This prompted her to request an explanation from the attorneys involved.
In response, Lunsford noted that the issue stemmed from his colleague Reeves’s use of ChatGPT without proper validation. Lunsford admitted the error, emphasizing that Reeves had utilized the AI to gather case citations without verifying their authenticity, leading to significant inaccuracies.
“We acknowledge this is unacceptable, embarrassing, and contradicts the respect we hold for the judicial system and all parties involved,” Lunsford stated, expressing sincere apologies for the lapse.
As a consequence of their actions, the judge removed all three lawyers from the case and mandated that they distribute her sanctions order to clients, judges, opposing counsels, and all attorneys in their Huntsville office. The ruling is also set to appear in a federal legal journal to ensure broader awareness of the misconduct.
“Creating false legal authority is a grave offense that necessitates strong repercussions,” Judge Manasco remarked, asserting that the situation warranted greater accountability than the common reprimands or minor fines typically issued for such violations.
She also criticized the Alabama Department of Corrections for retaining the attorneys despite their misconduct. “Their continued employment raises questions about the effectiveness of these sanctions when the attorney’s client, a state agency, is aware of the unethical behavior and chooses to maintain their representation.”
Reeves admitted his oversight in not verifying the AI-generated citations, stating, “I relied on the generated output without checking its validity. I deeply regret this lapse in diligence and the judgment that led to it.”
Judge Manasco highlighted Lunsford’s attempts to avoid the initial hearings regarding the sanctions, stating that such practices cannot be the standard among litigators, especially experienced ones, in federal court.
Observers say this incident underscores the growing ethical dilemmas surrounding the use of AI in legal settings, prompting discussions about the necessity for rigorous scrutiny and accountability in legal documentation. “This cannot be how litigators, particularly seasoned ones, practice in federal court or run their teams,” Manasco concluded.
This article was automatically written by Open AI and may contain inaccuracies. Any requests for removal, retraction, or correction can be sent to contact@publiclawlibrary.org.