HHS Will Use AI to Monitor State Audit Reports for Healthcare Fraud
The Department of Health and Human Services announced Thursday that it will use ChatGPT and other AI tools to analyze audit reports from all 50 states on an ongoing basis. The move aims to catch fraud in federal health programs and reduce government spending.
Gustav Chiarello, the assistant secretary for financial resources leading the program, said the department receives thousands of audit reports annually that go largely unreviewed. "It's classic big government: Everyone files an audit and it lands with a thud and no one does anything about it," he said. "Here, with AI, we're able to dig into it."
The initiative will scan audits from HHS-funded programs, including state Medicaid programs, federal research grants, addiction services, and other areas. States, local governments, nonprofits, and higher education institutions that spend at least $1 million in federal money yearly are required to submit annual audits.
How the Program Works
HHS said it has already sent letters to governors and treasurers in all 50 states notifying them of the new initiative. Recipients that fail to file required reports or resolve identified problems could lose funding.
Chiarello acknowledged that AI tools make mistakes and can carry biases, but said officials are evaluating public audit reports rather than uncovering new information. He framed the tools as a way to make grantees better stewards of federal dollars.
The Trump administration's anti-fraud task force, led by Vice President JD Vance, has promoted similar AI-based efforts across multiple programs, including Medicare, Medicaid, and student loan applications.
Concerns About Fairness and Accuracy
Critics have raised questions about the administration's fraud efforts, noting that most have targeted Democratic states. On at least one occasion, HHS acknowledged making a major error in data used to justify a Medicaid fraud investigation in New York.
Rob Weissman, co-president of Public Citizen, said the administration cannot be trusted to use AI tools fairly. "The AI is kind of beside the point when you assess what their actual objectives are, rather than what they pretend they are," he said.
Chiarello said he has contacted counterparts at other federal departments to encourage them to adopt similar AI monitoring systems. "It would be fairly easy for the other agencies to use our technology and jump on it," he said.
For government professionals managing federal compliance and audit functions, understanding AI for Government applications will become increasingly relevant as these tools expand across federal agencies.
Your membership also unlocks:
