The fact that artificial intelligence tools make it easier to fill out forms and prepare complaints is leading to significant increases in demands directed at public services. In the United Kingdom, complaints submitted to the housing ombudsman rose from 2,600 in 2022 to just over 7,000 last year, since ChatGPT began to be used. Complaints to the US Consumer Financial Protection Bureau (CFPB) also increased 5x over the same period. Similar rises were seen in court filings in Brazil and parliamentary petitions in Germany.
Researcher Chris Schmitz examines this trend, which he calls “agentic flooding,” in a paper to be presented at the AI Ethics and Society conference next month. Assessing 84 potential cases of such flooding across 11 jurisdictions, Schmitz says there is broad evidence that artificial intelligence tools are changing how people interact with public services. For methodological reasons, the study does not definitively say that the increase was directly caused by artificial intelligence. However, in most cases, submissions had remained flat before 2022, then accelerated and rose as artificial intelligence spread, and in most of them the increase has not yet slowed.
While some of the new submissions may be deliberate, Schmitz says most are legitimate requests made by real people. By reducing the “administrative burden” that previously made it difficult to submit applications, artificial intelligence may make it easier for people to access the social benefits or services to which they are entitled. The development could therefore create an opportunity to redesign public services for the artificial intelligence era.
Schmitz says online systems in public services can be used with an artificial intelligence assistant. Although he acknowledges that a resource problem similar to that faced by bug bounty programs, whose workloads increased last year due to low-quality reports generated by artificial intelligence, could arise, he emphasizes that most public applications contain valid requests. In his assessment, this marks the beginning of a period in which processes need to be reconsidered.
Why it matters
This development requires public institutions not only to process more applications, but also to address the administrative barriers to accessing applications and the evaluation processes together. Enabling people who previously struggled to prepare applications to join the system could have a positive effect on access to social benefits and services; however, the increased volume also expands the burden on institutions of distinguishing legitimate requests from deliberate or low-quality applications. While the findings are based on 84 cases across 11 jurisdictions, causality between the increase and the use of artificial intelligence has not been established. The open question, therefore, is how public systems will manage a lasting increase in the number of applications without excluding those entitled to benefits or wasting resources inefficiently.