AI agents are driving a surge in public-service requests

Complaints, petitions and applications have climbed in several countries since ChatGPT spread, raising concerns about workload and access.

As AI tools make it easier to fill out forms, draft complaints and submit applications, public services in several countries are seeing a sharp rise in requests. The pattern is already visible in the UK, the US, Brazil and Germany, where filings have jumped since ChatGPT became widely available. According to IT-PUB News, the shift matters because it could either overload public systems with extra work or help people who previously gave up on applying at all.

Complaints and petitions are climbing across countries

The clearest example comes from the UK, where complaints to the housing ombudsman have more than doubled since ChatGPT was introduced. The number rose from 2,600 in 2022 to just over 7,000 last year.

Similar increases have been recorded elsewhere. The US Consumer Financial Protection Bureau saw complaints grow fivefold over the same period. Researchers also found comparable jumps in Brazilian judicial petitions and German parliamentary petitions.

What is drawing attention here is the breadth of the pattern. It is not limited to one country or one kind of public service. It appears across welfare applications, judicial appeals and other online systems that can be used with help from an AI assistant.

Chris Schmitz describes the trend as agentic flooding

Chris Schmitz is tracking the rise as part of a broader pattern he calls “agentic flooding.” In a paper due to be presented next month at the AI Ethics and Society conference, he examines 84 potential flooding cases across 11 jurisdictions.

The research does not claim that AI is directly causing every surge. But the paper says most of the examples follow the same basic pattern: filings were roughly flat before 2022, then began rising faster as AI tools spread. In most of the cases studied, that growth has not slowed.

That makes the trend look like more than a brief spike. If it continues, public services may have to process far more requests without any matching increase in staff or budget.

AI is lowering the effort needed to apply

Schmitz says the increase is not hard to explain. AI systems are becoming more capable and easier to use, which means people can complete forms or draft complaints with much less effort than before.

He described a shift from carefully piecing together information and prompting earlier versions of ChatGPT to something much simpler — for example, taking a photo of a letter with a Claude app and getting a usable response in one step.

That matters because many people may have had valid claims but never filed them. In policy terms, this is often described as administrative burden: the effort needed to navigate forms, gather documents and finish a process that can feel intimidating or time-consuming.

From that angle, AI is not just creating more requests. It is also lowering the barrier for people who already had a right to ask for help.

Agencies may face the same workload strain seen elsewhere

The growing volume is being compared with what bug-bounty programs experienced last year, when companies were flooded with low-quality security reports written with the help of large language models. Those reports rarely pointed to serious vulnerabilities, but they still had to be reviewed, consuming time and resources.

Public services could run into a similar problem if request volumes keep rising. Even when applications are legitimate, a fivefold increase still means more work for the same agencies, often without extra funding or staff.

Schmitz says the comparison has limits. In the bug-bounty case, the main issue was a flood of poor submissions. In the public-service cases he studied, most of the new applications appear to come from real people with genuine claims.

“The vast majority of cases we find are people who are entitled to claim for something, claiming for that thing,” he told TechCrunch.

That distinction matters. The challenge is not simply filtering out spam. Public systems may also need to handle more demand from people who can now apply more easily.

Governments are weighing overload against better access

The harder question is how governments and agencies should respond. Some of the new filings are clearly adversarial, while others reflect legitimate applicants using AI to submit claims they might otherwise have abandoned.

That leaves a real tension. Agencies may see the trend as AI-generated pressure on already stretched services. At the same time, the same tools may help people get past paperwork, confusion and delays that have long kept them away from public support.

Schmitz argues that this could be a chance to rethink how services are designed. If AI can make applications easier, he suggests, the public sector may need to redesign its processes for an AI era rather than treat the rise in filings only as a nuisance.

He said there is “a good version” of AI use in these systems, where people are genuinely helped rather than blocked by complexity. In his view, the moment could push governments to reconsider how these services work from the ground up.

The broader point is straightforward: AI is not only changing chatbots and office software. It is also changing the flow of requests into public institutions, bringing a mix of wider access and heavier strain on systems that were not built for this level of automated assistance.