AIBRIEF
Phillips v. Parlade treats an AI written order as a 'judicial act' anyway, foreclosing a federal damages suit and pushing accountability to appeal and bar channels.
Phillips v. Parlade, decided last Wednesday (Aug. 12) in the District of Nevada, is the first reported federal ruling testing judicial immunity against an allegation that a state-court judge issued a ruling "wholly" through AI. U.S. District Judge Gloria Navarro dismissed the §1983 civil suit, holding that absolute immunity covers judicial acts "even for error, malice, or excess of authority." (Volokh Conspiracy / Reason)
The plaintiff argued that 100% AI delegation strips the order of "discretionary human thought," removing it from the "judicial act" category and placing the judge in the narrow "clear absence of all jurisdiction" exception. The court disagreed, applying the four-factor judicial-act test (normal judicial function, in chambers, centered on a pending case, arising from official confrontation) and finding each factor satisfied. The plaintiff cited no case law for the exception.
The ruling is narrow. The federal court did not adjudicate whether AI was actually used. The state-court judge has not admitted the allegation, and the underlying merits remain live. What is foreclosed is the federal civil-damages path.
Practical accountability still lives in three channels: a direct appeal in the underlying case, a grievance to the Nevada Council on Judicial Discipline, and a complaint to the Nevada State Bar, where sanctions can range from censure to license suspension. Community discussion has mapped these same remedies. For a litigant facing a judge who might lean on AI, the lever was never a federal damages suit. It was the appellate and discipline record, and that record is still open.