Deloitte's Center for Government Insights published guidance on how US federal and state health agencies could use AI systems that plan and carry out multistep tasks to detect fraud earlier and prepare investigation evidence faster. It cites government estimates of federal fraud losses in the hundreds of billions of dollars a year and a cross-border banking pilot that doubled detection of known fraud. The recommendations are prospective, and the authors stress that reviewers must verify every suggestion.
What changed
Health agency fraud control largely runs after payment, with provider risk re-checked at revalidation two to five years apart.
What it unlocks
Continuous monitoring of provider licensing and ownership changes, and plain-language querying of billing and policy data to test new fraud hypotheses.
- $233bn-$521bn est. annual federal fraud losses
- fraud estimate covers FY2018-2022
- 2x fraud detection in SWIFT pilot
- 13 banks in SWIFT data pilot
What you need to act on it
- documented agency procedures and policy knowledge before scaling
- human review of AI-generated leads
- data-sharing authority and privacy safeguards
Sources