As AI assistants have become widely available, many public services internationally have recorded sharp rises in applications, complaints and petitions. Researcher Chris Schmitz describes this phenomenon as “agentic flooding” and has documented patterns across multiple jurisdictions.
Scope of the research
In a paper to be presented at the AI Ethics and Society conference, Chris Schmitz examined 84 cases from 11 jurisdictions. The cases include housing ombudsman complaints, consumer financial complaints, judicial petitions and parliamentary petitions. Schmitz reports that most of these submissions were roughly flat before 2022, and then began rising rapidly as AI technologies diffused; in most instances the growth has not yet slowed.
Selected figures cited in the research:
- United Kingdom: complaints to the housing ombudsman rose from about 2,600 in 2022 to just over 7,000 in 2023, more than doubling.
- United States: the Consumer Financial Protection Bureau (CFPB) saw a fivefold increase in complaints over the same period.
- Similar jumps were observed in Brazilian judicial petitions and German parliamentary petitions.
For methodological reasons Schmitz stops short of asserting direct causation between AI and the surge, but he finds a consistent arc across most of the 84 cases: stability before 2022, then accelerating growth as AI use expands.
Spam, opportunistic filings or real claims?
The rise in volume raises two concerns. One is that public service inboxes could be burdened with low-value or adversarial filings. This mirrors last year’s experience in some bug-bounty programs, where companies were inundated with low-quality reports generated by large language models; those reports rarely identified major security issues but still consumed staff time to triage.
The other concern — and what Schmitz emphasizes — is that many of the new applicants are real people asserting legitimate claims. He told TechCrunch that “the vast majority of cases we find are people who are entitled to claim for something, claiming for that thing.” In other words, AI may be enabling filings that previously failed to materialize because the administrative work involved was too burdensome.
Schmitz notes that improvements in AI capabilities and accessibility have made it progressively easier for users to prepare submissions. Tasks that once required careful prompt-crafting and manual compilation of documents can now be handled by pasting a letter or photographing a document in an AI app and receiving a usable response quickly.
Implications for public services
Short-term, public agencies may face significant capacity challenges: more applicants with the same budgets and staffing levels. Long-term, however, the shift also presents an opportunity to redesign administrative processes for an AI-enabled environment.
Schmitz argues that rather than treating all new filings as spam, policymakers and service providers should consider how to build the ‘‘good’’ version of AI-assisted services — systems that reliably help users prepare accurate claims while protecting against abuse. Doing so will require both addressing immediate triage and verification needs and thinking strategically about how services are structured.
Next steps
Schmitz’s dataset is available alongside his paper. His analysis suggests the increase in demand will likely continue for years, and he calls for detailed design work to define and implement better public service processes that harness AI’s capabilities while managing the risks and resource implications.
