The Era of ‘Agentic Flooding’: How Generative AI Is Overwhelming Bureaucracy and Redefining Public Services

Executive Overview

Across the globe, public administration frameworks are confronting an unprecedented operational crisis driven by consumer-facing artificial intelligence. As generative AI models make drafting legal petitions, filing regulatory complaints, and navigating bureaucratic forms virtually frictionless, public services are experiencing exponential surges in citizen submissions. This emerging phenomenon—termed "agentic flooding"—is rapidly transforming the interface between citizens and the state.

In the United Kingdom, housing complaints submitted to the official ombudsman more than doubled within a two-year window. In the United States, consumer disputes filed with federal financial regulators grew fivefold. Similar spikes in official filings are destabilizing judicial systems in South America and parliamentary petition channels in Western Europe.

While public institutions historically designed their workflows around the assumption that citizen friction serves as an informal filter, the ubiquity of large language models (LLMs) like OpenAI’s ChatGPT and Anthropic’s Claude has dismantled that barrier overnight.

┌────────────────────────────────────────────────────────────────────────┐
│                   THE MECHANICS OF AGENTIC FLOODING                    │
├────────────────────────────────────────────────────────────────────────┤
│                                                                        │
│  [ Legacy Paradigm ]                                                   │
│  High Administrative Burden ──> Citizen Abandonment ──> Low Filings    │
│                                                                        │
│  [ Post-2022 AI Paradigm ]                                             │
│  Multimodal AI Tools ───────> Zero Friction ──────────> 5x Application │
│                                                         Volume Surge   │
└────────────────────────────────────────────────────────────────────────┘

A landmark study led by technology researcher Chris Schmitz tracks this trend across 84 distinct cases in 11 jurisdictions. The findings indicate that rather than acting as a wave of malicious automated spam, "agentic flooding" primarily represents legitimate citizens exercising rights they previously abandoned due to daunting "administrative burden." Consequently, public sector leadership faces a critical inflection point: treat the influx as an operational bottleneck to be defended against, or leverage this wave to fundamentally restructure public administration for the artificial intelligence era.

Detailed Chronology: The Administrative AI Surge (2022–2026)

The systemic shift in how citizens interact with state mechanisms did not occur in isolation; it correlates directly with the release cycles and interface evolution of consumer generative AI tools.

       PRE-2022                  LATE 2022-2023                   2024-PRESENT
┌─────────────────────┐      ┌─────────────────────┐      ┌───────────────────────────┐
│  Baseline Filings   │      │  Text-Based AI      │      │ Multimodal & Agentic      │
│                     │      │                     │      │                           │
│ • High friction     │ ────>│ • ChatGPT-3.5 launch│ ────>│ • Claude 3.5 & GPT-4o     │
│ • Manual paperwork  │      │ • Prompting needed  │      │ • Photo/PDF instant draft │
│ • Flat petition     │      │ • Initial filing    │      │ • Sustained exponential   │
│   growth rates      │      │   volume surges     │      │   volume growth           │
└─────────────────────┘      └─────────────────────┘      └───────────────────────────┘

Phase 1: The Pre-2022 Baseline

Prior to November 2022, citizen-initiated petitions, administrative appeals, and regulatory complaints grew at predictable, linear rates across most developed nations. Bureaucratic channels relied on the intrinsic complexity of application forms—often requiring precise legal jargon, physical documentation, and hours of manual compilation—to keep submission volumes within manageable operational capacities.

Phase 2: The Text-Based Influx (Late 2022–2023)

The public deployment of ChatGPT (GPT-3.5) marked the initial inflection point. Early adopters recognized that conversational LLMs could translate complex human grievances into structured, professional, and bureaucratically aligned text.

• United Kingdom: Housing Ombudsman complaint volumes, which had remained relatively stable at approximately 2,600 cases in 2022, began accelerating rapidly, reaching over 7,000 filings by the end of the following year.
• United States: The Consumer Financial Protection Bureau (CFPB) recorded early signs of elevated dispute submissions as consumers utilized AI tools to challenge credit bureau entries and banking fees.

Phase 3: Multimodal Frictionlessness and Agentic Diffusion (2024–2026)

As AI models evolved from text-only prompts to multimodal platforms capable of parsing images, PDFs, and unstructured documents (such as Claude 3.5 and GPT-4o), the technical threshold to generate complex claims vanished entirely. Instead of spending hours engineering precise prompts, citizens could simply capture a photo of a tenancy dispute notice, eviction warning, or utility bill, instructing the application to formulate a formal legal filing.

According to Schmitz’s dataset, submission volumes across public services have not plateaued. Instead, the growth curve continues on an upward trajectory across multiple global regions, indicating that generative tools are becoming permanent intermediaries in public life.

Supporting Context & Metrics: Cross-Jurisdictional Analysis

To analyze the structural impact of AI on public administration, Chris Schmitz examined 84 distinct cases of potential agency flooding across 11 major jurisdictions. The research paper, prepared for the AI Ethics and Society conference, establishes a clear quantitative baseline: across public services utilizing digital intake portals, submission rates remained flat prior to 2022, only to experience sharp, compound growth alongside consumer AI adoption.

Key Global Metrics

                       SURGE IN COMPLAINT FILINGS
     8,000 ─────────┐
                    │                                   7,000+
     6,000 ─────────┤                                 ┌────────┐
                    │                                 │        │
     4,000 ─────────┤                                 │        │
                    │            2,600                │        │
     2,000 ─────────┤          ┌────────┐             │        │
                    │          │        │             │        │
         0 ─────────┴──────────┴────────┴─────────────┴────────┴────────
                             UK Housing              UK Housing
                          Ombudsman (2022)        Ombudsman (Recent)

The Bug-Bounty Parallel: Noise vs. Intent

To comprehend the institutional strain caused by agentic flooding, technology analysts point to an instructive parallel in the private sector: cybersecurity bug-bounty platforms. Throughout 2024 and 2025, corporate security teams were submerged under an overwhelming volume of LLM-generated vulnerability reports—often referred to as "AI slop."

Because generative models frequently hallucinated security flaws or flagged trivial code anomalies, security teams spent hundreds of human hours vetting useless submissions. This created severe operational exhaustion without yielding meaningful security improvements.

┌────────────────────────────────────────────────────────────────────────┐
│               BUG-BOUNTY SLOP VS. PUBLIC AGENCY FLOODING               │
├────────────────────────────────────────────────────────────────────────┤
│                                                                        │
│  CYBERSECURITY BUG BOUNTIES                                            │
│  [ LLM Outputs ] ──> Hallucinations / "Slop" ──> High Noise, Low Value │
│                                                                        │
│  PUBLIC SERVICE AGENCIES                                               │
│  [ LLM Outputs ] ──> Legitimate Claims ───────> Low Noise, High Value  │
│                                                                        │
└────────────────────────────────────────────────────────────────────────┘

However, Schmitz’s empirical dataset reveals a critical distinction between commercial bug bounties and public service filings:

1. High Legitimacy: Unlike security vulnerability spam, the overwhelming majority of AI-assisted public filings originate from real citizens with legal claims.
2. Entitlement Realization: Applicants using AI to request welfare assistance, submit housing complaints, or appeal tax assessments are typically entitled to those benefits under existing law.
3. Targeted Submissions: The surge represents functional democratic participation rather than adversarial automated network attacks.

Official Statements & Policy Frameworks

The core dilemma facing public sector executives is distinguishing between adversarial system abuse and genuine civic engagement enabled by technology.

In an interview regarding his findings, Chris Schmitz emphasized that the evolution of AI model capabilities directly parallels the uptick in citizen compliance filings:

"People are finding out that this is something one can do, and incrementally, it is just getting easier to do it… Before, it might have been a question of a lot of dragging context together and prompting ChatGPT 3.5 very precisely; it may now be a question of just pasting or taking a photo of a letter with your Claude app and getting a pretty good response in one shot."

Addressing the critical question of whether public agencies are simply facing an influx of automated spam, Schmitz clarified:

"The vast majority of cases we find are people who are entitled to claim for something, claiming for that thing."

The "Administrative Burden" Paradox

In political science and public administration literature, the term administrative burden defines the learning, compliance, and psychological costs that states impose on citizens seeking access to public services. For decades, governments have relied on administrative friction—intentionally or unintentionally—to regulate demand and manage constrained operational budgets.

When generative AI tools absorb these compliance costs on behalf of the citizen, the hidden demand for public services becomes immediately visible.

┌────────────────────────────────────────────────────────────────────────┐
│                   THE ADMINISTRATIVE BURDEN TRIAD                      │
├────────────────────────────────────────────────────────────────────────┤
│                                                                        │
│    1. Learning Costs    ──> Understanding complex legal codes          │
│    2. Compliance Costs  ──> Filling multi-page forms & gathering proof   │
│    3. Psychological Costs ──> Navigating daunting bureaucratic systems │
│                                                                        │
│    * Generative AI tools mitigate all three costs simultaneously. *   │
│                                                                        │
└────────────────────────────────────────────────────────────────────────┘

Schmitz argues that this operational disruption should be recognized as a catalyst for institutional modernization rather than a administrative failure:

"A big part of making AI go well is being able to detail out what the good version of things looks like. And anyone who’s ever used ChatGPT to do their tax return knows that there’s a good version here where you’re being helped. This could be the moment to say, ‘We need to rethink pretty much everything about how this process looks.’"

Future Outlook: Rebuilding Public Bureaucracy for the AI Era

The rapid onset of agentic flooding exposes a fundamental asymmetry in modern governance: citizens are deploying 21st-century AI tools to interact with administrative architectures designed in the 20th century. As budget-constrained public agencies attempt to process five times their historic filing volumes without corresponding staffing increases, traditional manual review processes are nearing total exhaustion.

┌────────────────────────────────────────────────────────────────────────┐
│            STRATEGIC ROADMAP FOR PUBLIC AGENCY MODERNIZATION           │
├────────────────────────────────────────────────────────────────────────┤
│                                                                        │
│   ┌───────────────────┐    ┌───────────────────┐    ┌──────────────┐   │
│   │ Automated Triage  │ ─> │ API-First Intake  │ ─> │ Human-in-the │   │
│   │ & AI Summaries    │    │ Platforms         │    │ -Loop Review │   │
│   └───────────────────┘    └───────────────────┘    └──────────────┘   │
│            │                        │                       │          │
│            ▼                        ▼                       ▼          │
│   Categorize incoming      Bypass dense PDF forms    Focus human staff │
│   petitions instantly      with structured data      on complex claims │
│                                                                        │
└────────────────────────────────────────────────────────────────────────┘

Policy experts and public sector strategists outline a four-stage roadmap for managing this systemic transformation:

1. Automated Intake and AI-to-AI Triage

To handle the volume of incoming filings, public agencies must deploy internal AI intake systems capable of summarizing, categorizing, and verifying AI-generated submissions. By establishing machine-readable standards for claims processing, agencies can automatically validate documentary evidence before routing files to human caseworkers.

2. Transition to API-First Public Infrastructure

Dense, multi-page PDF forms and complex web portals were designed for human manual entry. As AI agents increasingly manage citizen administrative tasks, governments can shift toward secure, authenticated application programming interfaces (APIs). Structured digital intake channels eliminate the need for citizens to generate verbose text documents simply to pass basic information to state databases.

3. Preserving Equitable Democratic Access

As agencies adopt anti-bot measures to manage portal traffic, care must be taken to avoid inadvertently disenfranchising non-technical or vulnerable demographics. Security implementations must distinguish between malicious denial-of-service attempts and legitimate citizens utilizing consumer AI assistants to assert their statutory rights.

4. Structural Reallocation of Civil Service Resources

By automating initial intake, document verification, and administrative triage, civil service personnel can transition away from repetitive administrative data entry. Staff resources can instead be directed toward high-value, empathetic case management, complex dispute resolution, and direct community outreach.

Conclusion

Agentic flooding represents a fundamental evolution in civil engagement. While the initial surge in AI-generated filings has strained legacy administrative infrastructure, it highlights a profound opportunity for civic structural reform. By embracing automated intake frameworks and redesigning public portals for an era of low-friction technology, governments can build more accessible, efficient, and responsive public institutions.