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

The study identifies “agentic flooding,” where large-language-model generated appeals cause surges in demand for government benefits, citing 84 probable cases across 11 jurisdictions. A risk matrix shows financially attractive yet complex services are most vulnerable, and while fee-based friction can curb flooding, it may compromise equitable access; the authors propose near-term mitigations that avoid this trade-off.