For decades, legal information providers have offered copies of standard contracts, powers of attorney, and other blank legal documents for everyday needs. But as helpful as they can be, no provider can claim to offer legal advice to any one customer who uses their forms.
Today, the advent of generative AI tools like ChatGPT has allowed users to conduct legal research, draft business documents, and create individualized pleadings purely through artificial intelligence. Self-represented litigants can now produce reams of sophisticated-looking filings without even buying specialized software.
According to the New York Times, judges and opposing counsel alike are struggling with the impact of an increase in pro se litigation, together with volumes of filing, fueled by AI tools.
Why and When People Go Without Lawyers
Representing oneself in legal proceedings is one of American law’s oldest rights. It is true that some people act pro se because they refuse to listen to attorneys, leading to the ancient joke: “A man who represents himself has a fool for a client.” However, that saying is both untrue and unfair for many pro se litigants, as many simply cannot pay attorneys’ fees or find anyone willing to represent them. Without the right knowledge of civil procedure or legal research, it does not take a “fool” to make a case-ending mistake; even attorneys have done that.
In the past, pro se litigants had to rely on templates — sometimes even in their own handwriting — and whatever legal research they could do in law libraries or internet searches. Before ChatGPT and other LLMs, many might have given up on their claims, regardless of their merits. Today, LLMs can produce pages and pages of work product, offering options for motion practice that a pro se litigant might not have dreamed of otherwise.
The Effect on Courts, Counsel, and Cases
A recent study reports a considerable rise in civil, non-prisoner pro se cases in U.S. courts, together with an increase in volume of filings in those cases, over 158 percent more docket entries than before AI. The study also claims that 18 percent of sampled complaints from 2026 test positive for AI use. According to the NYT, this study is not yet peer-reviewed. However, if the study’s findings are accurate, it reflects what judges and attorneys are experiencing.
Even when a pro se litigant pays nothing to create their filings, they have a cost. Judges, opposing parties, and court staff must handle them: docketing them, taking the appropriate actions, reviewing their arguments, and drafting responses. All this takes time and money for everyone involved, and those resources are wasted on filings that cite “AI hallucinations” or simply drag out the proceedings. (In fairness, attorneys have also allowed AI mistakes to slip into court filings.)
Future Possibilities for AI Legal Advice
States, courts, and private parties are pushing back as they can. Some courts have issued standing orders governing the use of AI. The NYT points to the U.S. District Court for the Middle District of Georgia, which cautions pro se litigants that the court can sanction them for inaccurate filings or those that “harass, cause unnecessary delay, or needlessly increase the cost of litigation.” One judge in the Northern District of Ohio has an order forbidding AI use entirely in matters he oversees in federal court.
The state of New York is considering legislation that would penalize AI companies for chatbots “working” in law, medicine, psychology, and other professional fields. A House bill called the CHATBOT Act would have the FTC regulate the use of AI chatbots and provide a private right of action, though its future is uncertain. One company has even sued OpenAI to demand compensation for a lengthy litigation which, it alleges, is fueled by ChatGPT’s advice to one unhappy woman. (OpenAI has since changed its terms of service and disavowed its use as professional advice, but it is still possible to make the chatbot produce the work anyway.)
Should laws and courts prevent AI’s use, or simply prevent its abuse? Attorneys and law firms use specialized legal AI software for tasks such as drafting, document review, and even appearing before the Supreme Court. It is possible that future pro se litigants could have more reliable AI tools with less potential for abuse. But even high-quality software simply cannot act as an attorney: a person who knows the local court rules, has experience in front of the judges, and can firmly and kindly tell a litigant no.
