Why Self-Represented Litigants Outnumber Lawyers in AI Hallucination Court Cases

Most coverage of AI hallucination sanctions focuses on lawyers, because a licensed attorney getting sanctioned is the more dramatic story. But, interestingly, in the largest running database of these cases, self-represented litigants — people with no attorney at all — actually outnumber lawyers by a wide margin. That split tells a different story than the one usually told.

Quick answer

Of the 2,042 court decisions tracked in Damien Charlotin’s AI hallucination database as of mid-September 2026, 1,174 involve a pro se (self-represented) litigant, compared with 812 involving a lawyer. The gap reflects something specific: AI tools have lowered the barrier to filing something in court at all, and that access comes with a verification gap nobody built a safety net for.

Why the split exists

A self-represented litigant using a general-purpose AI tool to draft a filing has no professional supervision structure, no CLE requirement pushing verification habits, and often no awareness that AI-generated legal citations can simply be invented. A lawyer using the same tool at least operates inside a system — bar rules, malpractice exposure, colleagues — that creates some pressure toward checking the output. That system doesn’t exist for someone representing themselves.

This isn’t a story about who’s more careless. It’s a story about where the safety net has holes. AI tools made it meaningfully easier for someone without legal training to produce something that looks like a real legal filing. They did not make it easier for that person to know the citations inside it were fabricated.

What happens differently for each group

• Pro se litigants most often receive warnings, admonishments, or have a brief struck — courts frequently note that pro se status doesn’t excuse the duty to verify, but sanctions tend toward correction rather than punishment on a first instance.
• Lawyers face the fuller range of consequences: monetary sanctions, CLE requirements, bar referrals, public reprimands, and in the more severe cases, suspension. The professional licensing structure that pro se litigants don’t have is exactly what creates the additional consequences for attorneys.

What this signals for the profession

Access to legal recourse has genuinely expanded — people who couldn’t afford representation can now draft something resembling a legal filing on their own. That’s not automatically bad. But it’s expanded without the verification infrastructure that makes a filing trustworthy, and courts are absorbing that gap case by case rather than through any systemic fix.

For law firms specifically, this split is also a quiet signal about the market: a growing pool of people are trying AI-assisted self-representation before — or instead of — hiring an attorney. Some of them will hit a wall (a case complex enough, a filing rejected enough times) that sends them looking for real representation. Firms with a strong, honest web presence are positioned to catch that moment; firms that aren’t easy to find or evaluate online simply lose that person to whichever competitor is.

FAQ

Does this mean pro se AI hallucination cases are less serious?
Not less serious to the underlying case — a fabricated citation can still get a filing struck or a claim dismissed. The sanctions are typically lighter because there’s no professional license to leverage as a consequence, not because the error matters less.

Are courts changing how they screen pro se filings because of this?
Some jurisdictions have adopted specific AI-use policies at the court level (Illinois and California’s Judicial Council among them), but this varies significantly by jurisdiction and isn’t yet a uniform practice.

Where does this data come from?
The AI Hallucination Cases Database maintained by legal researcher Damien Charlotin, updated near-daily and licensed for reuse with attribution.

---

Source: AI Hallucination Cases Database, Damien Charlotin — updated daily, licensed CC BY 4.0.