As generative AI makes it easier to write documents and file complaints, the number of public service applications and submissions is surging in countries around the world. Borrowing the phrasing of researcher Chris Schmitz, it is a phenomenon dubbed Agentic Flooding.

TechCrunch reported on Sept. 10 that in Britain, Housing Ombudsman complaints rose from 2,600 in 2022 to more than 7,000 last year after the adoption of ChatGPT. Complaints to the U.S. Consumer Financial Protection Bureau (CFPB) rose fivefold over the same period. Similar increases were seen in judicial petitions in Brazil and parliamentary petitions in Germany.

In a paper Schmitz will present at an AI ethics and society conference in October, he reviewed 84 cases across 11 jurisdictions. The subjects ranged from welfare applications to formal judicial appeals, but all shared a common feature: online services accessible to AI assistants.

The paper did not conclude, from a methodological standpoint, that AI directly affected the increase in complaints. But in most cases, submissions were generally stagnant until before 2022, then the pace of increases accelerated after the spread of AI. In many cases, that rise has not yet slowed, TechCrunch reported.

Public agencies may therefore have to process far more applications with the same budgets.

Bug bounty services that accept reports of security vulnerabilities are facing a similar situation. A flood of low-quality reports produced by large language models has poured in, and companies have had to review every submission.

Public services appear to be different, though. Schmitz sees most cases as qualified people asserting real rights. That suggests people who had previously given up exercising their rights because the application process was too burdensome are using AI to regain access to the system, TechCrunch reported.

Schmitz sees public services as needing to go beyond merely responding to higher volumes and redesign the entire application process for the AI era.