Chris Schmitz Publishes Study Linking AI Tools to Surging Public-Service Requests

Public agencies may be facing a new kind of demand shock: people can use AI to turn intimidating forms, complaints and appeals into submissions they can complete. Researcher Chris Schmitz has published an analysis of 84 potential cases across 11 jurisdictions that connects this shift to sharp growth in public-service requests, while stopping short of claiming AI caused it.

The stakes are not limited to a crowded inbox. A person who once abandoned a benefit claim or official complaint because the process was too difficult may now get help assembling it. But the same tools can send far more requests into public systems whose staff must still assess each case.

A rise across very different systems

Schmitz calls the pattern “agentic flooding.” His cases span online services from welfare applications to judicial appeals. In most, submissions were roughly flat before 2022, then rose at an accelerating pace as AI tools became more widely available; most had not yet shown a slowdown.

The examples stretch beyond complaints. The analysis also points to increases in Brazilian judicial petitions and German parliamentary petitions. That breadth is central to Schmitz’s argument: this is not simply a single agency changing its rules or a single public controversy generating responses.

From paperwork to prompt

The proposed mechanism is practical rather than mysterious. AI can help someone collect context, interpret an official letter and draft a response. Schmitz said the task may have shifted from carefully prompting an earlier chatbot to pasting in, or photographing, a letter and receiving a usable response in one step.

What the extra volume could represent

• Legitimate applicants who previously gave up because filing a claim was too burdensome.
• Adversarial submissions, which the analysis recognizes alongside ordinary claims.
• A heavier assessment workload for agencies, regardless of how a request was prepared.

Correlation is not a verdict

The paper does not establish that AI directly caused the increases. Its evidence is a recurring timeline: demand was stable, then grew as AI spread. That leaves room for other forces behind any individual surge, and means the analysis should not be read as proof that every additional filing was machine-assisted.

It also complicates the instinct to label high-volume AI-assisted requests as junk. Schmitz told TechCrunch that the vast majority of cases he found involved people entitled to make a claim. If that assessment holds, filtering out volume alone could exclude people who have gained access to processes designed around difficult paperwork.

The service-design question

The immediate policy tension is between capacity and access. Public systems may need ways to handle a larger flow of requests without treating every AI-assisted filing as suspect. Schmitz’s more optimistic reading is that reducing administrative burden could be a reason to rethink how claims processes work, not merely a reason to defend their old limits.