Voters Oppose AI Fraud Detection When It Could Deny Eligible Benefits

In an October poll of 2,527 registered voters, 53% said they would not support AI tools that hunt fraud in federal programs if those tools might mistakenly deny benefits to qualified individuals, while 37% remained in favor. The earlier August poll, which did not mention the risk of wrongful denials, showed 60% support for AI use. The latest questioning added a Medicaid example and highlighted potential errors, prompting a notable drop in approval, especially among Democrats, independents, and women.

Analysts note that the pattern mirrors findings from GAO reports on real-world programs where tighter controls can create barriers for legitimate claimants. Critics argue that removing human oversight could increase false positives and reduce accountability, underscoring the need for balanced fraud-prevention measures.