Justice by Chatbot? Why More Americans are Turning to Artificial Intelligence to Fight Their Own Legal Battles

For years, access to justice in the United States has often depended on one thing: money. Those unable to afford lawyers in civil disputes have frequently been left with little choice but to abandon their claims or attempt to navigate the legal system alone.

Now, artificial intelligence is beginning to alter that landscape, according to a Reuters report.


Across the United States, a growing number of self-represented litigants are relying on AI tools such as ChatGPT, Grok and Claude to draft lawsuits, prepare motions, research legal arguments and respond to opposing counsel. What was once considered an intimidating and inaccessible process is increasingly being approached with the help of generative AI systems that can produce legal documents within seconds.

The trend is becoming visible across federal courts. Research by economics doctoral students Joshua Levy of the University of Southern California and Anand Shah of the Massachusetts Institute of Technology found that self-represented litigants accounted for nearly 17 per cent of all federal civil court cases in the fiscal year ending September 30, 2025. That marks a sharp increase from the long-term average of around 11 per cent. The federal appellate court system separately recorded a 9 per cent rise in appeals filed by litigants without lawyers during fiscal 2025.

The figures point to a broader transformation underway in the American legal system, where AI is increasingly filling gaps left by costly legal services and declining legal-aid support.

Unlike criminal defendants, who are constitutionally entitled to public defenders if they cannot afford representation, individuals involved in civil disputes must generally pay for their own legal assistance. Although nonprofit organisations and government-funded legal-aid groups provide support in some cases, demand has long outpaced available resources.

A 2022 report by the Legal Services Corporation estimated that low-income Americans failed to receive adequate or any legal assistance for 92 per cent of their civil legal problems. Against that backdrop, AI tools are emerging as a low-cost alternative for people who would otherwise struggle to access the courts.

One such litigant is Las Vegas resident Nicole Silverberg, who turned to AI after failing to find a lawyer willing to take her case on a contingency basis. Unable to afford upfront legal fees, she used paid versions of ChatGPT and Grok to help pursue a federal lawsuit against her former landlord, alleging retaliatory eviction linked to water contamination complaints and withheld rent.

According to Silverberg, the AI systems assisted with procedural guidance, including filing motions and identifying weaknesses in opposing arguments. But her experience also exposed the dangers of relying too heavily on chatbot-generated legal work.

At one stage, a judge criticised her for submitting hallucinated case citations in a preliminary injunction filing and warned of possible sanctions. The incident forced her to scrutinise AI-generated content more carefully and highlighted a growing concern among judges and legal professionals: while AI may increase access to courts, it can also generate inaccurate legal authorities, misleading advice and excessively lengthy filings.

Judicial concerns are mounting nationwide. According to a database maintained by HEC Paris senior research fellow Damien Charlotin, US judges have reprimanded lawyers and self-represented litigants nearly 1,000 times in recent years over improper AI use in court filings.

Several judges have begun issuing direct warnings. In Georgia, US District Judge Marc Treadwell recently ordered a self-represented litigant to disclose any AI use and verify the accuracy of court submissions, citing a noticeable increase in AI-assisted filings from unrepresented parties unfamiliar with legal obligations.

Legal academics and court observers say the problem is not merely about factual inaccuracies. They argue that AI is also reshaping the structure and workload of litigation itself.

Levy and Shah found that pro se cases now contain significantly heavier dockets, with more filings, arguments and procedural activity than before the AI boom. Their research showed docket entries per case rose by 38 per cent in mid-2025 compared with pre-AI averages.

Although the researchers stopped short of claiming definitive proof that AI alone caused the increase, they concluded that the overall pattern strongly suggests a connection.

That surge is creating new pressure on already burdened courts.

Courtroom5, an AI-powered online platform designed for self-represented litigants, says technology is helping more people push their claims beyond the initial filing stage. Its chief executive, Sonja Ebron, believes AI is enabling litigants to prepare documents sufficiently well to survive early procedural hurdles that might previously have blocked them from proceeding.

At the same time, critics warn that accessibility does not necessarily translate into stronger legal cases.

Jeffrey Cohen, an associate professor at Boston College Law School who supervises students assisting self-represented litigants in federal court, says AI often encourages users to produce unnecessarily long and overly detailed filings instead of concise legal arguments.

He also warns that chatbots tend to encourage litigation by affirming potential claims without properly assessing legal barriers such as limitation periods or evidentiary weaknesses. Some clients, he says, are becoming reluctant to accept professional legal advice when it conflicts with AI-generated responses.

The debate reflects a deeper tension now emerging inside the American legal system: whether AI should be viewed primarily as a democratising force that expands access to justice, or as a disruptive technology that risks overwhelming courts with weak, inaccurate or procedurally flawed litigation.

Major AI companies are already moving aggressively into the legal sector. This week, Anthropic announced partnerships aimed at expanding AI-powered legal tools for individuals and small businesses unable to afford traditional legal representation, alongside new legal features for its Claude platform.

The rapid adoption of AI in law suggests the technology is unlikely to remain a fringe experiment. Instead, it may become a permanent feature of civil litigation — particularly for people priced out of conventional legal services.

For litigants like Silverberg, the technology has already altered life trajectories. Her experience representing herself in multiple legal proceedings inspired her to self-publish a book on pro se litigation and enrol in Purdue Global Law School’s online programme with the goal of eventually practising law in California.

Her lawsuit against her landlord remains ongoing after lengthy efforts to serve multiple defendants properly. But the broader significance of her case lies beyond the dispute itself.

It illustrates how AI is beginning to redraw the boundaries of legal access in America — empowering ordinary individuals to enter courtrooms once considered financially out of reach, while simultaneously forcing judges, lawyers and policymakers to confront the consequences of a justice system increasingly shaped by machines.