Lawyers Keep Citing AI-Hallucinated Cases Despite Repeated Warnings
Key Takeaways
- ▸Lawyers have been sanctioned for citing AI-generated fake cases, with at least one attorney continuing the practice after being formally warned
- ▸A curated database documents 1,400+ court rulings addressing AI errors, with new cases appearing at a steady rate of 350–400 per quarter
- ▸AI hallucinations have also affected journalists, developers, researchers, and government consultants across multiple professional sectors
Summary
Courts across the United States have sanctioned multiple attorneys for filing legal briefs containing false citations generated by AI systems, with judges noting that some lawyers continued citing nonexistent cases even after being warned about the practice. A database maintained by HEC Paris researcher Damien Charlotin documents over 1,400 cases where courts have addressed AI errors since 2023, with new cases entering the record at a rate of 350–400 per quarter as of late 2025.
The problem extends far beyond the courtroom. Journalists, software developers, academic researchers, and government consultants have all been caught relying on AI-fabricated information, from misattributed quotes in published books to hallucinated code and false citations in research. The pattern suggests a systemic issue: people continue to trust AI outputs despite widespread awareness that these systems can and do generate plausible-sounding falsehoods—a phenomenon researchers attribute to a unique form of bias toward machine-generated information.
Psychological research reveals the roots of this misplaced confidence. A February 2025 study found that people who view AI favorably are more likely to accept incorrect AI guidance than incorrect guidance attributed to humans, suggesting that AI systems inspire a particular kind of trust that overrides normal skepticism. Experts warn that as AI becomes more embedded in professional workflows, the consequences of this unchecked reliance will escalate beyond dismissed legal appeals and attorney fines.
- Psychological studies show people exhibit unusual trust in AI outputs compared to human guidance, even when both are equally unreliable
- The fundamental issue is not AI literacy but behavioral: people know AI can hallucinate but continue to rely on outputs without verification
Editorial Opinion
The crisis of AI hallucinations in professional settings reveals a troubling gap between knowledge and behavior. While lawyers, researchers, and journalists are well aware that AI systems fabricate information, many continue to use these tools without adequate verification—a failure that carries real consequences, from dismissed legal appeals to compromised scholarship. As AI becomes more prevalent in knowledge work, organizations and professions must move beyond awareness campaigns and implement structural safeguards: mandatory fact-checking processes, integration with verification tools, and accountability measures for inadequate due diligence.
