Complaints to the United Kingdom’s housing ombudsman more than doubled between 2022 and last year, rising from 2,600 to just over 7,000, according to figures cited by researcher Chris Schmitz. Over the same period, the United States’ Consumer Financial Protection Bureau saw its complaint volume grow roughly fivefold. Neither agency changed its remit, its staffing model, or its publicity strategy. What changed, Schmitz argues, is the ease with which people can now file a complaint at all.
Schmitz is tracking the phenomenon as part of a broader trend he calls agentic flooding — the rapid rise in applications, petitions, and filings that occurs once AI assistants make administrative tasks trivial to complete. His paper, set to be presented next month at the AI Ethics and Society conference, examines 84 potential cases of flooding across 11 jurisdictions, spanning welfare applications, official judicial appeals, and everything in between.
Also read: OpenAI Pauses New ChatGPT Pro Sign-Ups as Astra Demand Strains Its Systems
A pattern that predates every specific AI product
The cases Schmitz examined share a consistent shape. Submissions in each jurisdiction were roughly flat before 2022, then began rising at an accelerating pace as AI technology diffused into everyday use. Crucially, most of the 84 cases had not yet seen that growth slow down, suggesting the trend is likely to continue for years.
Alongside the UK housing ombudsman and the CFPB, Schmitz documented similar jumps in Brazilian judicial petitions and German parliamentary petitions. His full dataset is hosted publicly for other researchers to examine. For methodological reasons, the paper stops short of declaring AI the direct cause of the surge — but the correlation across jurisdictions and service types is difficult to attribute to coincidence.
Also read: Meta's AI Agent Muse Hits No. 2 on US App Store With 83,000 Downloads
Schmitz points to a simple mechanism. “People are finding out that this is something one can do, and incrementally, it is just getting easier to do it,” he told TechCrunch. “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.”
The bug bounty parallel — and where it breaks down
The surge has a close analogue in the security industry. Last year, bug-bounty programs reported being overwhelmed by low-quality vulnerability reports generated by large language models. The reports rarely contained genuine security issues, yet companies were still obligated to triage each one, consuming significant engineering resources.
Public services face a superficially similar problem: the same budget, five times the applicants. But Schmitz’s data points to a different conclusion. “The vast majority of cases we find are people who are entitled to claim for something, claiming for that thing,” he told TechCrunch.
If those people were not claiming benefits or filing complaints before, it was often because the process itself was too forbidding. Policy researchers call this administrative burden — the friction of forms, deadlines, and documentation that quietly excludes people from services they are legally entitled to use. AI tools, in Schmitz’s framing, are lifting some of that burden for the first time.
An opportunity rather than a crisis
That distinction matters for how governments respond. A service that treats every new filing as potential spam will invest in filtering, verification, and rejection. A service that treats the surge as evidence of previously suppressed demand might instead redesign its intake process entirely.
Schmitz leans toward the second approach. “A big part of making AI go well is being able to detail out what the good version of things looks like,” he said. “And anyone who’s ever used ChatGPT to do the 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.'”
The practical work, however, has barely begun. Most of the jurisdictions in Schmitz’s dataset have not yet adjusted their processes, staffing, or digital infrastructure to account for AI-assisted filing. Omudbsman offices and consumer bureaus continue to operate on intake systems designed for a pre-2022 baseline.
What happens next depends partly on whether governments treat the increase as a resource problem or a design problem. The paper’s presentation next month is likely to draw attention from civil servants and regulators already grappling with the same question — and Schmitz’s data suggests the volume will keep rising regardless of which answer they choose.
Frequently Asked Questions
What is agentic flooding?
Agentic flooding is a term coined by researcher Chris Schmitz to describe the rapid growth in applications, complaints, and petitions submitted to public services as AI tools make it easier to complete administrative tasks that were previously too burdensome to pursue.
Are AI-generated public service requests a form of spam?
Schmitz’s research suggests the opposite is more common: the vast majority of new filings come from people with legitimate claims who previously were deterred by the administrative burden of applying. A smaller share of submissions are clearly adversarial.
How much have complaints increased in the UK and US?
Complaints to the UK housing ombudsman rose from 2,600 in 2022 to just over 7,000 last year, while the US Consumer Financial Protection Bureau saw complaint volumes grow roughly fivefold over the same period.
What should governments do about the surge in AI-assisted applications?
Schmitz argues the increase is an opportunity to redesign public services to be more AI-friendly, rather than treating the additional volume purely as a resource problem to be filtered out.
Be the first to comment