A recent California Court of Appeal decision in Noland v. Land of the Free, L.P. has broken new ground in AI-related legal ethics. The court sanctioned lawyer Amir Mostafavi to the tune of $10,000. It referred him to the state bar after finding that nearly all the case citations and quotations in his appellate briefs—generated with ChatGPT and similar AI tools—were fabricated. Out of 23 quotations in the opening brief, 21 were fake, with many more appearing in the reply brief.
The court described the appeal as “unremarkable” but published its judgment to address, for the first time in California, the problem of fabricated AI citations. It emphasised that such misconduct wasted judicial resources and justified sanctions under California’s procedural rules.
The Twist: Opposing Counsel Denied Fees
Unusually, while the court sanctioned Mostafavi, it declined to award attorneys’ fees or costs to opposing counsel. The reason? They had neither detected the fake citations nor alerted the court. The judges suggested that if opposing counsel had identified the issue, fee-shifting might have been appropriate.
Sanctions raise a new professional responsibility question: Do lawyers now have a duty not only to check their own authorities but also to scrutinise an opponent’s citations for possible AI fabrications?
Broader Implications
Traditionally, lawyers could assume opposing citations were genuine, even if they still needed to verify accuracy and relevance. However, in the AI era, courts may expect heightened diligence, as fabricated cases can appear plausible and slip into filings undetected.
The Noland ruling hints at an evolving standard of competence:
- Spotting AI hallucinations could become part of a lawyer’s professional obligations.
- Failing to detect obvious fakes might limit entitlement to costs, even if the opponent is sanctioned.
- Rewarding vigilance may encourage lawyers to police the misuse of AI in court filings actively.
Looking Ahead
While Noland stops short of creating a clear rule, it signals that courts may soon expect both sides to play a role in detecting AI hallucinations. As fabricated citations become more sophisticated, clearer standards of diligence Will be needed. For now, the case underscores a new reality: in the age of generative AI, lawyers can no longer take an opponent’s citations at face value.
Source: LawSites
Leave a Reply