Public services are increasingly strained by LLM-written appeals for benefits

[Submitted connected 17 Aug 2026 (v1), past revised 19 Aug 2026 (this version, v2)]

View PDF HTML (experimental)

Abstract:AI agents are making it easier for the nationalist to interact pinch government, specified arsenic by helping them use for benefits, understand analyzable policies, and make their opinions heard. Although improving work accessibility is beneficial, immoderate resulting surges successful request could strain unprepared authorities services. We word specified surges agentic flooding of authorities services ("flooding") and supply 3 contributions. First, based connected a collected dataset of 84 imaginable cases of flooding crossed 11 jurisdictions, we posit that flooding is apt occurring wide today, mostly done ample connection models (LLMs) generating matter cheaply. Second, we measure what services are astir exposed to flooding. We create a consequence matrix to analyse a service's exposure, and propose that near-term consequence is highest for financially attractive, but analyzable services. Finally, we representation imaginable authorities responses to flooding. Precedent suggests these responses will apt beryllium capable to extremity astir cases of flooding, but the fastest to deploy - friction-inducing measures for illustration fees - often waste and acquisition disconnected equitable entree to nationalist services. Accordingly, we adjacent by recommending near-term actions that whitethorn let governments to mitigate flooding without invoking this trade-off.