Debt relief companies have spent the past two years facing a steady rise in demand letters and self-filed lawsuits from unrepresented consumers, particularly under the Telephone Consumer Protection Act (TCPA). When Shipkevich PLLC attorneys hosted the December 2025 webinar, The Pro Se Playbook: How Unrepresented Plaintiffs and Pre-Litigation Demands Are Reshaping the TCPA Landscape, we identified this wave as one of the fastest growing risks in the industry. Two new empirical studies now document just how large the wave has become, and both find substantial evidence consistent with generative artificial intelligence lowering the barrier to filing suit. Both also point to the same conclusion recently highlighted by Law360, which is that the spike in filings has not been matched by success in the courtroom.[1]
The numbers are striking. Self-represented litigants initiated roughly 23,000 federal cases in 2022 and approximately 41,000 in 2025, with the pro se share of filings rising from a long-standing level of about 11 percent to 16.8 percent in fiscal year 2025.[2] More than 18 percent of sampled 2026 complaints were flagged as containing AI-generated text, up from essentially zero before 2023,[3] and the growth is concentrated in comparatively formulaic civil claims rather than complex, attorney-intensive litigation, a pattern that should concern industries that routinely defend standardized statutory claims, including TCPA defendants.[4] Yet the results tell a different story. Complaints bearing indicators of AI drafting are dismissed more often than those without them and show no win-rate advantage,[5] and pro se plaintiffs overall still prevail in fewer than 1 percent of cases, essentially unchanged from the years before AI tools became widely available.[6] Courts are increasingly warning self-represented litigants that reliance on AI does not excuse inaccurate filings and can result in sanctions.[7]
The trap for debt relief companies is treating the low win rate as good news and moving on. A claim that fails still has to be defended, and AI now lets a single filer generate a polished complaint and respond to a motion to dismiss with little time or expense, driving up defense costs in cases that might once never have been filed.[8] The approach we outlined in the webinar is built for this environment. Keep consent and call records complete, challenge defective pleadings early, watch for template language and serial filer patterns, and enforce arbitration provisions where available. AI appears to be changing the volume and cost of pro se litigation without meaningfully improving outcomes, and companies that prepare for that increased volume will spend far less than those that wait for it.
Notes
[1] Jack Karp, AI Drives Spike In Pro Se Filings But Not Courtroom Success, Law360 Pulse (Aug. 13, 2026).
[2] Anand V. Shah and Joshua Y. Levy, Access to Justice in the Age of AI: Evidence from U.S. Federal Courts (Mar. 2026), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6766859.
[3] Shah and Levy, supra note 2 (AI text detection analysis of 1,600 randomly sampled federal civil complaints, 2019 through 2026, using the Pangram Labs detector).
[4] Shah and Levy, supra note 2.
[5] Or Cohen-Sasson, Artificial Access to Justice: AI and the Surge in Pro Se Litigation (June 2026), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6864398 (reporting dismissal rates of 61.1 percent for AI-flagged complaints versus 53.6 percent for unflagged complaints, drawn from a linked sample of non-form pro se complaints in selected case categories; the difference in plaintiff win rates was not statistically significant).
[6] Shah and Levy, supra note 2.
[7] AI-Powered Pro Se Litigation Is Flooding Federal Courts, What Businesses Need to Know, ArentFox Schiff (June 2026).
[8] Big Law Grapples With AI-Fueled Pro Se Surge, Rising Legal Costs, Bloomberg Law (Mar. 2026).