Oregon Supreme Court Strikes Down Cases Over AI-Generated Fake Citations
The Oregon Supreme Court has officially addressed the rise of artificial intelligence in legal filings, marking the first time the state’s high court has ruled on the use of fabricated citations. In two separate cases, the court struck down filings containing AI-generated falsehoods, highlighting a growing trend where lawyers and individuals representing themselves submit non-existent case law and fake quotes to the judiciary.
In an Oregon district court, a lawyer was previously fined $110,000 for using AI to submit filings, which stands as the highest penalty recorded to date for this type of infraction.
Why AI Fabrications Burden the Court
The significance of these rulings lies in the strain that AI-generated misinformation places on judicial resources. Chief Justice Meagan A. Flynn stated that when the court must spend significant time and effort addressing fabricated legal arguments, it inevitably comes at the expense of other pending cases. Researchers estimate that more than 1,000 cases nationwide now contain inaccuracies linked to the use of AI tools.
According to Ankur Doshi, General Counsel of the Oregon State Bar, the issue is compounded by the fact that many individuals do not realize the AI is generating false information. Beyond the time required for courts to verify these claims, opposing counsel often faces additional work, which can lead to higher attorney fees for the parties involved.
Samantha Carter notes that while AI may appear to be a convenient shortcut for legal research, the reliance on automated tools without human verification creates a systemic risk. The primary stake here is the integrity of the judicial record; when the court is forced to act as a fact-checker for machine-generated errors, it slows the administration of justice for everyone else.
Case Details and Judicial Consequences
In one instance, the Oregon Supreme Court dismissed a petition for a writ of mandamus after the plaintiffs submitted documents using LegalAI that contained fake quotes and cases that do not exist. Despite being asked to resubmit, the plaintiffs filed a new declaration less than 12 hours later that still included citations to at least four non-existent cases.
In a second case, the respondent submitted a document referencing legal arguments that could not be verified within Oregon case law. The court responded by issuing a $500 fine and granting the respondent an opportunity to resubmit the document correctly. This follows a national pattern of sanctions, including a $3,000 fine issued last summer in federal court against lawyers representing the chief executive of MyPillow for similar AI-driven errors.
What May Happen Next
Legal professionals and self-represented litigants are likely to face increased scrutiny when submitting research documentation. Given the Oregon Supreme Court’s firm stance, it is possible that courts will implement stricter verification requirements or mandate disclosures regarding the use of AI tools in document preparation. Observers might expect that further sanctions could be applied if the trend of submitting unverified, AI-generated citations continues to compound the workload of the state’s judiciary.
Frequently Asked Questions
What was the specific outcome for the plaintiffs in the writ of mandamus case?
The Oregon Supreme Court dismissed the petition after the plaintiffs submitted documents containing fake quotes and non-existent case citations, even after being given an opportunity to correct their work.
How much was the fine in the second case addressed by the court?
The respondent was fined $500 and was permitted to resubmit their document without the unverified legal arguments.
Why is the use of AI in legal research considered problematic by the Oregon State Bar?
According to General Counsel Ankur Doshi, AI fabrications create significant extra work for the court and opposing counsel, as they must dedicate resources to verifying the authenticity of every citation, which can ultimately increase legal costs for the opposing side.
How do you think the legal profession should adapt its training to address the risks posed by AI-generated research?