Policy
AI Use in Court Filings Jumps to 34% in California Counties
Stanford research reveals sharp adoption divide between self-represented litigants and attorneys filing civil complaints.
Key takeaways
- Over one-third of civil filings in Santa Clara and San Francisco counties now show signs of AI use, up sharply from 2024.
- Self-represented litigants in Santa Clara County show 72% AI adoption rates, compared to 34% for attorney-filed complaints.
- Employment and business cases demonstrate the highest AI usage rates among case types examined.
- AI content appears most frequently in statements of facts and causes of action sections of court filings.
- Stanford researchers conclude AI in legal practice is now a question of implementation rather than adoption.
More than one-third of civil lawsuits filed in two major California counties this year show evidence of artificial intelligence use, according to new research from Stanford University released Thursday.
The study examined court filings in Santa Clara County and San Francisco County, excluding debt collection cases. Researchers found that AI adoption has accelerated dramatically, particularly among people representing themselves in court.
Self-represented litigants drive adoption
The research reveals a striking gap between pro se litigants and those represented by attorneys. In Santa Clara County, the proportion of self-represented complaints showing AI indicators surged from 19% in 2024 to 72% in 2026. Attorney-filed complaints showed more modest growth, rising from 5% to 34% over the same period.
This pattern suggests AI tools are lowering barriers to legal system access for individuals who cannot afford representation, though the research does not assess the quality or success rate of these filings.
Employment and business cases lead usage
AI adoption varies significantly by case type. Employment disputes and business-related complaints in California showed the highest rates of AI use, according to the Stanford researchers. The technology appears most frequently in specific sections of filings—particularly statements of facts and causes of action, where litigants must construct legal narratives and arguments.
Why it matters
This research quantifies a shift that courts and legal professionals have sensed anecdotally. The finding that nearly three-quarters of self-represented complaints now contain AI elements has immediate implications for court administrators, judges, and opposing counsel. Courts may need to develop new protocols for identifying and addressing AI-generated content, particularly when it contains hallucinated citations or legal errors. The technology's rapid adoption also raises questions about access to justice—whether AI is democratizing legal processes or creating new risks for unrepresented parties who may not understand the limitations of these tools.
"The sharp rise we document here corroborates that AI in the law is a question of how, not whether," the Stanford researchers wrote in their report.
The findings were first reported by KQED and the ABA Journal. The research focused on California counties but likely reflects broader national trends as generative AI tools become more accessible and sophisticated.