The number of 'self-represented lawsuits' without lawyers, facilitated by AI, is skyrocketing, increasing the burden on courts.



Researchers at MIT and the University of Southern California have reported that the number of self-represented plaintiffs in the United States has surged since the proliferation of generative AI. The researchers argue that while AI is expanding access to justice, it is also significantly increasing the administrative burden on courts.

Access to Justice in the Age of AI:Evidence from US Federal Courts
(PDF file) https://avshah1.github.io/assets/pdf/papers/pro-se/Pro_Se_Automation.pdf

The AI justice gap solution is slowly turning into an existential paperwork nightmare for US federal courts
https://the-decoder.com/the-ai-justice-gap-solution-is-slowly-turning-into-an-existential-paperwork-nightmare-for-us-federal-courts/

AI Legal Filing Surge Threatens to Swamp Courts
https://winbuzzer.com/2026/05/25/mit-researcher-warns-ai-filing-surge-could-swamp-courts-xcxwbn/

The study analyzed approximately 4.5 million federal civil lawsuits filed by non-incarcerated individuals and approximately 46 million records from the federal court records system, PACER, from fiscal year 2005 to fiscal year 2026. It found that the proportion of self-represented plaintiffs remained stable at around 11% for many years, but rose to approximately 17% by fiscal year 2025.

The increase in self-representation in litigation was concentrated not in areas requiring specialized knowledge, such as patents or securities fraud, but rather in areas where the format of document creation is relatively standardized, such as civil rights lawsuits, consumer credit disputes, and seizure-related lawsuits. Researchers speculate that this is because AI can easily assist in creating complaints and petitions if they are in a well-structured format.



The problem is that a well-formed legal document doesn't necessarily contain a valid argument. For example, in the case of Donald Sove in Minnesota, a complaint drafted using ChatGPT and Claude, along with 50 supplementary documents and a case law analysis, was submitted, but the court ultimately determined that it did not present a clear claim.

In these cases, the fact that handwritten complaints were dismissed relatively quickly is now being transformed into professional-looking documents by AI is causing more verification work for the courts. Since each document must be received, recorded, and made public by staff, it takes considerable effort for judges to make a decision. According to the report, the number of record items within 180 days of an incident will increase by 158% by 2025 compared to the average before the widespread adoption of AI, meaning that courts are spending a tremendous amount of time processing records such as applications, responses, orders, and setting of hearing dates.



The use of AI is also evident in the documents themselves. A research team extracted 1,600 federal civil complaints from 2019 to 2026 and examined them using Pangram's AI document detection. They found that the percentage of complaints containing AI-generated text had risen from 1% in 2023 to 18% by the beginning of 2026.

However, AI-powered legal assistance cannot be dismissed outright. Researchers point out that 'for people who cannot afford legal fees, AI could be a means to access legal information and format documents.'



However, while the cost of initiating a lawsuit may decrease, the processing costs for the courts and opposing parties will increase. While AI broadens access to judicial participation, it has also revealed the limitations of a judicial system that relies on human judgment and administrative processing.

in AI, Posted by log1i_yk