The Download: AI-generated lawsuits and virtual power plants for data centers

Federal magistrate judge Maritza Braswell of Colorado has witnessed a seismic shift in courtroom documentation over the past eighteen months. Her chambers, like those of judges across the United States, have experienced a more than doubling of pro se filings—legal documents submitted by individuals without legal representation—since the beginning of 2023. The driving force behind this surge is artificial intelligence, specifically large language models and chatbots that can now draft legal motions, pleadings, and supporting documentation with minimal human intervention. While this technological democratization appears on its surface to expand access to the justice system for economically disadvantaged litigants who cannot afford traditional legal counsel, emerging data suggests a troubling paradox: increased filing volume has not translated into improved success rates for those leveraging AI assistance. The judicial system now confronts a crisis of both volume and quality, with courts struggling to process an unprecedented number of filings while questioning the fundamental competence and reliability of the AI-generated arguments they contain. This development raises urgent questions about the appropriate role of artificial intelligence within legal institutions and the regulatory frameworks needed to govern its use.

The proliferation of AI-generated litigation reflects a broader tension in technology adoption that has intensified throughout 2023 and into 2024. Access to justice has long been acknowledged as a critical challenge in democratic societies, with prohibitive legal costs creating barriers for low-income and middle-class Americans seeking redress for genuine grievances. The emergence of generative AI tools promised to overcome this structural inequality by providing sophisticated legal writing assistance to those who could never afford a lawyer's billable hours. However, the rapid deployment of these tools without corresponding safeguards or training mechanisms has created a new category of systemic problem: courts are now inundated with technically voluminous but legally deficient filings that consume judicial resources without advancing legitimate legal claims. The judicial system, already burdened by case backlogs and resource constraints, now faces the additional challenge of distinguishing between legitimate pro se filings and those that reflect algorithmic hallucinations or fundamental misapplications of legal doctrine. This moment represents a critical juncture in how society manages the integration of artificial intelligence into institutions that require precision, accountability, and professional judgment.

The quantitative dimensions of this shift reveal the scale of change occurring within the American legal system. Judge Braswell has documented that pro se filings in her jurisdiction have more than doubled relative to pre-2023 baseline levels, a figure that aligns with anecdotal reports from judges in other federal districts. Simultaneously, the nature of these filings has changed markedly, with judges increasingly noting that documents bear the stylistic hallmarks and structural patterns characteristic of large language model outputs rather than attorney-drafted pleadings or self-represented litigant work. The correlation is unmistakable: the availability of free or low-cost AI legal writing tools has lowered the threshold for filing, enabling individuals to submit documentation that would previously have been deemed too technically deficient to warrant court processing. What remains unstudied but critically important is the actual merit rate of these AI-assisted filings. Early indications suggest that despite the increase in pro se filings, the percentage that succeed in achieving the filer's legal objectives has not increased proportionally, implying that volume expansion has not been accompanied by quality improvement or increased litigation success for vulnerable populations who lack traditional legal counsel.

For legal professionals, courts, and individuals navigating the justice system, this development presents immediate and concrete challenges that extend beyond theoretical concerns about AI governance. Courts must now expend judicial resources screening AI-generated filings for basic legal competence before substantive merits review can occur, effectively creating a new administrative burden on an already overburdened system. Judges confront questions about professional responsibility when AI-generated filings contain legal arguments that are patently frivolous or based on hallucinated case law that does not exist—a phenomenon documented in legal proceedings involving attorneys who relied on ChatGPT-generated citations. The practical problem intensifies when considering that individuals filing pro se with AI assistance may not possess the legal knowledge to identify when an AI tool has generated fundamentally flawed arguments, leaving them to discover only through adverse rulings that their AI advisor had steered them toward losing positions. More broadly, the economics of litigation have been disrupted in ways that disadvantage traditional legal practice; defendants and opposing parties now must respond to AI-generated filings that, while potentially meritless, cannot be simply ignored by the legal system regardless of their quality. This creates asymmetric litigation costs where the sheer volume of filings generates work for opposing counsel and judicial officers even when those filings lack legal foundation.

The emergence of AI-generated litigation reflects a deeper pattern evident across multiple institutional domains: the implementation of artificial intelligence tools without corresponding governance frameworks, professional standards, or accountability mechanisms. Similar dynamics are visible in healthcare, where AI diagnostic tools are being deployed before their reliability has been rigorously established; in hiring, where algorithmic decision-making systems perpetuate or amplify existing biases; and in content moderation, where AI systems make consequential decisions about speech with limited human oversight. The legal system's struggle with AI-generated lawsuits is emblematic of a broader societal challenge: technology companies and enthusiasts have moved with extraordinary speed to deploy AI tools across critical institutional functions, but governance structures, professional ethics frameworks, and regulatory oversight have lagged substantially behind. The judiciary's implicit assumption in these cases is that if AI can write something that resembles legal documentation, it should be processed through normal legal channels—but this assumption fails to account for quality, reliability, and the systemic costs of processing low-quality filings at scale. Lawmakers and judges are increasingly grappling with fundamental questions: should there be professional licensing requirements for individuals using AI to draft legal documents for others? Who bears liability when an AI system generates legally deficient advice that harms the person relying upon it? Should courts implement screening mechanisms specifically designed to identify and manage AI-generated filings differently from traditional pro se submissions?

The trajectory ahead will be determined by decisions made within the next twelve to eighteen months by courts, bar associations, and state legislatures. Federal courts and state judiciaries are beginning to establish local rules addressing AI use in litigation, with several districts already requiring disclosure when AI has been used in document preparation. The American Bar Association continues to develop ethical guidelines for attorney use of AI, but these standards do not address the pro se population that is generating the bulk of AI-assisted filings. Observers should monitor developments at the state bar association level, particularly whether licensing bodies will create certification requirements for AI-assisted legal document preparation or establish standards for non-lawyers using these tools. Additionally, the evolution of this issue will likely shape legislative responses; bills addressing AI liability and professional accountability are being drafted in multiple states, and the outcomes of these regulatory efforts will determine whether AI-generated litigation becomes an integrated feature of American justice or a cautionary tale about uncontrolled technology deployment. Judge Braswell and her colleagues are currently managing this transition through discretionary judgment and case-by-case decision-making, but without systematic policy frameworks, the problem will likely intensify as AI tools become more sophisticated and their accessibility expands further.