By Sami Mahroum
Published: September 11, 2026
ABU DHABI — Governments worldwide are increasingly turning to artificial intelligence to process mountains of bureaucratic data, optimize resource allocation, and streamline public services. Yet, a fundamental paradox is emerging at the intersection of automation and statecraft: as machines grow exponentially more capable, human attention remains strictly finite. Far from eliminating administrative gridlock, the widespread deployment of AI in the public sector risks simply shifting the bottleneck to where it matters most—human oversight.
Nowhere is this friction more palpable than in the justice system. Since the public debut of generative AI tools like ChatGPT in late 2022, civil institutions have struggled to keep pace with a tidal wave of machine-generated text. The legal system, built on centuries of deliberate precedent and meticulous human review, is suddenly on the front lines of an automated paper chase.
Main Facts
The integration of artificial intelligence into public administration and legal frameworks has transitioned from a theoretical future-of-work discussion into an urgent operational crisis. The core issue lies in asymmetric generation: AI can produce complex legal filings, policy feedback, regulatory appeals, and citizen grievances in mere seconds. However, evaluating, verifying, and adjudicating those documents still requires the cognitive bandwidth of human judges, clerks, and regulators.
In the legal sector, this asymmetry has manifested as an unprecedented surge in pro se litigation—lawsuits filed by individuals representing themselves without formal legal counsel. Armed with sophisticated language models, everyday citizens and opportunistic litigants can draft dense, legally formatted complaints that mimic the output of seasoned attorneys.
While proponents argue that AI democratizes access to justice by lowering financial barriers to the legal system, it simultaneously overwhelms court dockets. Judges and administrative staff are now forced to parse through mountains of filings, many of which contain fabricated case law (hallucinations), irrelevant arguments, or superficial legal jargon, threatening to grind the wheels of justice to a halt.
Chronology of the AI Legal Surge
To understand how rapidly generative AI has disrupted public administration and the courts, it is necessary to examine the trajectory of the phenomenon over the past several years:
- November 2022: OpenAI publicly releases ChatGPT, igniting a global technological arms race and introducing consumer-grade generative AI to the general public.
- Throughout 2023: Early adopters begin experimenting with AI tools to draft legal documents, academic papers, and bureaucratic correspondence. Research data indicates that the share of federal civil complaints in the United States containing machine-written text sits at a modest 1%.
- Throughout 2024: As large language models (LLMs) become more integrated into daily software suites and web browsers, the barrier to entry for generating sophisticated text drops to near zero. Courts begin noticing erratic filings containing hallucinated citations and bizarre legal reasoning.
- Throughout 2025: The volume of AI-assisted litigation reaches critical mass. Courts in various jurisdictions report significant backlogs driven by unrepresented litigants utilizing AI to flood the system with paperwork. Statistical analyses reveal that the baseline of pro se federal lawsuits climbs sharply.
- September 2026: Current empirical data reveals a systemic transformation. The baseline share of federal lawsuits filed without a lawyer jumps to nearly 17%, while the share of federal civil complaints featuring machine-written text skyrockets to 18%. Governments and international legal bodies grapple with the reality that human oversight capacity has officially been outpaced.
Supporting Data and Empirical Evidence
The subjective feeling among legal professionals that courts are being flooded with AI output is heavily supported by recent quantitative research. A landmark study tracking federal litigation trends provides stark metrics on the velocity of this technological shift.
Historically, the share of federal lawsuits filed by pro se litigants—individuals navigating the court system without a licensed attorney—remained relatively stable. Between 2005 and 2022, the historical average hovered around 11%. However, following the widespread availability of generative AI tools, that figure climbed to 16.8% by 2025.
Even more dramatic is the metric tracking the composition of the filings themselves. The same research indicates that the share of federal civil complaints containing machine-written text stood at just 1% in 2023. Today, that number has surged to 18%—an eighteen-fold increase in less than three years.
This data illustrates a profound structural reality: the friction of writing has been effectively eliminated. Where drafting a comprehensive civil complaint once required hours of professional legal labor or financial investment, it now requires only a well-crafted prompt. Yet, the friction of reading, verifying, and judging that text remains anchored to human cognitive limits. A judge who could previously review a manageable number of complex filings per day cannot proportionately increase their reading speed, comprehension, or fact-checking capacity to match the exponential output of automated generators.
Official Responses and Regulatory Struggles
Governments, judicial councils, and administrative agencies are scrambling to formulate coherent responses to the AI-generated paperwork crisis. The approaches taken thus far range from punitive measures to technological counter-measures.
Judicial Sanctions and Mandatory Disclosures
Across multiple federal and state jurisdictions, judges have begun implementing strict standing orders regarding the use of generative AI. Several federal judges now require attorneys—and in some cases, unrepresented litigants—to explicitly certify that any AI tool used in drafting documents has been thoroughly verified for accuracy. When lawyers have failed to do so, resulting in court filings with fabricated case citations, judges have levied severe financial sanctions and mandatory reporting to state bar associations.
Administrative Safeguards
Beyond the courtroom, public administration agencies are deploying their own AI filters to screen incoming public comments, regulatory petitions, and benefit applications. Ironically, governments are fighting AI with AI. Automated triage systems are being designed to flag machine-generated submissions, assess their validity, and prioritize human review for high-risk files. However, this creates a secondary arms race, as citizens utilize more advanced models designed specifically to bypass bureaucratic detection filters.
Policy Gaps and Institutional Resistance
Despite these localized interventions, a cohesive global or national framework remains elusive. Legislative bodies move deliberately, often taking years to pass comprehensive technology regulations, whereas AI capabilities evolve on a monthly basis. Public sector institutions face severe budgetary and hiring constraints, making it difficult to attract the technical talent required to audit and manage sophisticated automated systems.
Implications for the Future of Governance and Justice
The collision of generative AI and public administration portends profound structural changes for how societies are governed, how laws are enforced, and how justice is administered.
The Illusion of Democratic Access
On the surface, the rise of AI-assisted pro se litigation appears to democratize the legal system, empowering ordinary citizens to challenge corporate entities or state agencies without paying exorbitant legal fees. However, this accessibility may be largely illusory. If the courts are flooded with low-quality, machine-generated noise, the actual quality of justice deteriorates for everyone. A system overwhelmed by unverified filings cannot effectively adjudicate legitimate grievances, ultimately harming the very citizens the technology was meant to empower.
Burnout and the Human Bottleneck
In public administration, the ultimate bottleneck is not computational power, server capacity, or data storage—it is human attention, judgment, and accountability. When AI tools generate thousands of automated responses, policy inquiries, or legal motions, they transfer the cognitive burden onto civil servants and judges. Without a corresponding expansion of the judiciary and civil service (an economically and politically unlikely prospect), public officials face unprecedented burnout. The risk is a hollowed-out administrative state where overworked officials rely increasingly on automated summaries to make life-altering decisions about citizens, stripping the human element from justice and governance.
Redefining Public Trust
Trust in public institutions relies heavily on the perception of deliberate, fair, and human-led processes. If citizens come to believe that courts and government agencies are merely rubber-stamping decisions made by algorithms—or conversely, that the system is entirely bogged down by automated spam—public confidence in the rule of law will erode.
As Abu Dhabi and other global innovation hubs look toward the next generation of smart governance, the lesson is clear. Technology can exponentially scale the production of bureaucracy, but it cannot manufacture wisdom. Navigating the AI era will require public institutions to recognize that efficiency is not merely a matter of processing more data, but of preserving the finite, invaluable space required for human thought, empathy, and deliberation.
