# A Federal Judge Ruled Judges Are Immune Even If AI Wrote Their Rulings > Source: > Published: 2026-08-20 14:28:16+00:00 *A federal judge has made the AI question in one Nevada case legally simple: even if a state judge relied wholly on artificial intelligence to write a ruling, judicial immunity still blocks a damages lawsuit against her.* U.S. District Judge Gloria Navarro of the District of Nevada dismissed Phillips v. Parladé with prejudice in an order filed on August 12, 2026, after plaintiff T. Matthew Phillips accused Nevada family court Judge Mari D. Parladé of turning over her decision-making duties to artificial intelligence. According to Reason's Volokh Conspiracy, which highlighted the order on August 17, Phillips argued that total reliance on AI meant the ruling was not really a judicial act at all. Navarro didn't need to decide whether the AI allegation was true. That part is simple. Issuing a ruling in a pending case is a normal judicial function, she found, and judicial immunity protects judges from civil liability for that work even when the work is alleged to be wrong, malicious, or beyond their authority. No trial was needed. That's a strange fact pattern to see in a federal order. But it produced a very conventional immunity ruling. ## The ruling was narrow Phillips tried to frame the case around human judgment. His objection said judicial immunity should not apply because Parladé allegedly relied wholly on AI, without any discretionary human thought, and therefore stepped outside the judicial role. Navarro rejected that move. The court said the challenged conduct was still judicial in nature because it involved a ruling in Phillips's own state court case, there was no allegation that the events took place outside chambers, and Phillips offered no authority showing a clear absence of jurisdiction. [Claude AI Suffers Widespread Outage Across All Its Models on August 18](https://startupfortune.com/claude-ai-suffers-widespread-outage-across-all-its-models-on-august-18/) Claude went down across every major model on August 18, with Downdetector logging more than 4,000 user reports within an hour and Anthropic's status page offering no root cause. The outage is the latest in a run of at least seven incidents over the prior six days, hitting Claude Chat, the mobile app and Claude Code. - [claude AI outage August 18 all models down](https://startupfortune.com/claude-ai-suffers-widespread-outage-across-all-its-models-on-august-18/) - [anthropic status page confirms widespread Claude errors today](https://startupfortune.com/claude-ai-suffers-widespread-outage-across-all-its-models-on-august-18/) That last phrase matters. Judicial immunity is not a reward for careful work. It is a shield attached to the function. A judge can make a bad ruling, issue a careless order, or exceed her authority and still be immune from a damages suit if she is acting as a judge in a case before her. The usual remedy is appeal, not a separate federal lawsuit against the judge. Reason's Eugene Volokh made the same point in plain terms: the federal court did not decide whether Phillips's allegation was accurate, but even assuming it was, this lawsuit could not go forward. You may hate that result if you think AI should never be allowed near judicial decision-making. The doctrine doesn't care. It asks what function the judge was performing. The earlier report and recommendation from Magistrate Judge Nancy Koppe had also pointed to the Rooker-Feldman doctrine, because Phillips wanted the federal court to void or expunge a state-court minute order dated August 15, 2023. Navarro declined to reach that issue once judicial immunity resolved the case. That door is shut. ## The discipline question remains The order does not bless AI-written rulings. It says a private federal damages action is the wrong weapon. That distinction is not legal hair-splitting for anyone who may end up before a judge using software to draft or summarize orders. If the concern is that a judge delegated too much to a tool, the remaining routes are state appellate review, mandamus-type relief, or judicial discipline. Nevada has already had a public warning about AI in court filings. That one was on the lawyers' side of the bench. LawSites reported in September 2025 that Washoe County District Judge David Hardy confronted two Cozen O'Connor lawyers after a brief in the CommNV and Uprise case contained AI-generated fictitious citations. Removal from the case, referral to the Nevada bar, $2,500 fines: those were the sanctions on the table, though Hardy suspended the fines if the lawyers completed an alternative education-focused remedy. That was lawyer misconduct. Phillips v. Parladé raises a harder question, because the allegation points at the judge's own order, not a lawyer's brief. Courts punish lawyers for fake cases easily enough: a signed filing gives them a clear hook. A judge's ruling is different. Judicial immunity does what it has always done here: it protects the act from a damages suit even before anyone reaches the messy factual question underneath. Frankly, that's where the discomfort sits. If a lawyer files a brief with hallucinated cases, the court can sanction the lawyer. If a litigant believes a judge let AI do the judging, the answer is slower and less direct. Appeal is slower. Discipline is uncertain. But a damages suit against the judge is not available just because the alleged tool changed. [OpenAI Disbands Its Preparedness Safety Team Ahead of a Blockbuster IPO](https://startupfortune.com/openai-disbands-its-preparedness-safety-team-ahead-of-a-blockbuster-ipo/) OpenAI has dissolved its Preparedness team, the unit built to catch whether frontier models could enable bioweapons or rogue self-hacking systems, folding its work into existing teams. It's the third safety-focused team the company has shut down in two years, and it comes alongside the exits of COO Brad Lightcap and ethics chief Chloé Bakalar as... - [openai disbands preparedness safety team before IPO](https://startupfortune.com/openai-disbands-its-preparedness-safety-team-ahead-of-a-blockbuster-ipo/) - [AI safety teams dissolved ahead of company valuation](https://startupfortune.com/openai-disbands-its-preparedness-safety-team-ahead-of-a-blockbuster-ipo/) Nevada's Code of Judicial Conduct already uses the familiar language of avoiding impropriety and the appearance of impropriety. Phillips alleged a mangled version of that phrase in the underlying state proceeding, but Navarro's order did not test judicial ethics against AI-assisted judging. It only decided immunity. The warning is plain: AI may force courts to write new rules for chambers, but it has not cracked one of the oldest shields judges have. **Also read:** [CoreWeave Posts a Blowout Quarter Right as AI Bubble Fears Peak](https://startupfortune.com/coreweave-posts-a-blowout-quarter-right-as-ai-bubble-fears-peak/) • [Binance Lets AI Agents Trade Your Crypto With No Cap on Losses](https://startupfortune.com/binance-lets-ai-agents-trade-your-crypto-with-no-cap-on-losses/) • [Cerebras Unveils CS-4 Chip It Claims Is 30 Times Faster Than Nvidia GPUs](https://startupfortune.com/cerebras-unveils-cs-4-chip-it-claims-is-30-times-faster-than-nvidia-gpus/)