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

AI-generated appeals for government benefits are overwhelming public services through a phenomenon called "agentic flooding," where large language models make it cheap and easy for people to submit applications, straining unprepared government agencies across multiple jurisdictions. The research identifies that financially attractive but complex services face the highest near-term risk, and while traditional government responses like fees can mitigate flooding, they risk reducing equitable access to public services. The authors recommend near-term mitigation strategies that address the problem without imposing barriers that disadvantage vulnerable populations.