According to a report, generative AI might be driving a spike in pro se filings, but it is not leading to success in the courtroom.

As discussed by Angela Delvecchio on Project Counsel Media (AI might be driving a spike in pro se filings, but it is not leading to success in the courtroom, available here), GenAI has triggered a massive surge in self-represented (pro se) federal court filings. You probably already know that. What you may not know is that it’s not making pro se litigants any more likely to win their cases.

In her article, Delvecchio covers a new paper about AI’s impact on federal court filings titled Access to Justice in the Age of AI: Evidence from U.S. Federal Courts from Joshua Y. Levy, an economics Ph.D. student at the University of Southern California and Massachusetts Institute of Technology Ph.D. student Anand V. Shah.

The 56-page paper is available here. Delvecchio describes it as “massive”, I would add “dense” as well.

For those not inclined to read such a large paper, Delvecchio thankfully summarizes several of the findings, as follows:

• Pro se litigants filed an average of about 23,000 new federal lawsuits each year between 2005 and 2022, the year ChatGPT was first released to the public. But pro se litigants filed: 27,370 suits in 2023, 31,478 in 2024, 41,490 in 2025 and 29,323 so far in 2026 (as of May).
• Levy and Shah found that non-prisoner, self-represented litigants accounted for about: 11% of civil federal court filings in the years between 2005 and 2022, 14.5% of those filings in 2024 and 16.8% of them in 2025.

AI is likely responsible for much of the increase in pro se filings because:

• The increase in pro se cases is most obvious in what Shah and Levy refer to as “simple” cases, such as lawsuits involving credit consumer disputes, civil rights violations, foreclosures and employment discrimination, rather than more “complex” cases like patent and product liability suits.
• A growing percentage of pro se filings bear the signature of AI-generated text, according to scholars. Levy and Shah ran a random sample of 1,600 federal civil complaints – 200 from each year between 2019 and 2026 – through the AI-text detection service provided by Pangram Labs. Findings: just 0.1% of the filings analyzed were found to contain AI-generated text during the period from 2019 to 2022, 3.5% of filings in 2024, 10.5% in 2025 and 18% in early 2026.

While I certainly know that AI-text detection is often not very accurate, these numbers are still compelling.

Is AI paying off for pro se parties? Not exactly. Interesting bits from the Levy/Shah study:

• Judges dismissed approximately 60% of pro se cases before the advent of AI and 63% of pro se cases in the years since AI was widely introduced
• Roughly 20% of pro se cases were settled before AI and about 23% are post-AI
• Self-represented litigants still win fewer than 1% of cases – just as they did before the AI era

Delvecchio also cites a second study – this one by the Miami Law & AI Lab at the University of Miami School of Law. Their research shows:

• 61.1% of complaints containing a signature of AI-generated text are dismissed, compared to 53.6% of complaints without an AI signature
• Suits filed using AI ended before the defendant answered the complaint 47.8% of the time, compared with 39% of complaints that weren’t flagged for AI use
• Plaintiffs won lawsuits filed using AI at a rate of 6.4%, while they won 8.4% of those cases filed without AI

So (by these numbers, at least), GenAI might be driving a spike in pro se filings, but it’s not leading to more success in the courtroom.

Delvecchio discusses much more in her article here, including an example of someone who turned to AI to draft and file their complaints and more meaning behind the numbers.

So, what do you think? Are you surprised using GenAI isn’t helping pro se parties win more cases? Please share any comments you might have or if you’d like to know more about a particular topic.

Image created using ChatGPT, using the term “a group of robots standing in line to file cases in a courthouse”.

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.