The US Department of Health and Human Services says it will use ChatGPT and other AI tools to analyse audit reports from all 50 states on an ongoing basis. Led by assistant secretary Gustav Chiarello, the programme aims to identify fraud risks and cost issues in federally funded health programmes, while critics warn about AI errors and bias.
International
-Sathish Raman
The US Department of Health and Human Services said on Thursday it was increasing its use of artificial intelligence. The agency said the plan would review how states and other recipients check federally funded health programmes. Officials said the goal was to reduce fraud risks and cut government costs. The department planned to use ChatGPT and other tools to read audit reports regularly.
The US Department of Health and Human Services says it will use ChatGPT and other AI tools to analyse audit reports from all 50 states on an ongoing basis. Led by assistant secretary Gustav Chiarello, the programme aims to identify fraud risks and cost issues in federally funded health programmes, while critics warn about AI errors and bias.
Gustav Chiarello, the assistant secretary for financial resources, said the system would scan audits from all 50 states. Chiarello said audits often get filed but receive little follow-up. “Its classic big government: Everyone files an audit and it lands with a thud and no one does anything about it,\” Chiarello said. \”Here, with AI, were able to dig into it.\”
HHS AI tools and federal health audit reports
Under federal rules, states, local governments, nonprofits and higher education institutions must file yearly audits. The requirement applies when an organisation spends at least USD 1 million in federal funds. Chiarello said the new work would cover audits tied to HHS-funded programmes. These included state Medicaid programmes and grants for research and addiction services.
Chiarello said the AI review would run on an ongoing basis. Officials said it would focus on identifying patterns and issues in long documents. The agency also warned that recipients must address problems found in reports. Groups that fail to file reports, or fix issues, could lose funding. HHS said it had informed governors and treasurers in every state.
HHS AI tools and broader use of generative AI
The new audit plan added to HHS work with generative AI in other areas. The department had used such tools to examine state Medicaid programmes. Officials also used AI to automate administrative work and edit text. Chiarello said the audit project aimed to improve oversight of federal dollars. The program was first reported by The Wall Street Journal.
Supporters said AI could help spot trends across large sets of documents. Critics warned such tools can be wrong and may produce biased results. Questions about errors have followed other government efforts using automated checks. In at least one case, the administration told The Associated Press it used incorrect data. That data supported a New York Medicaid fraud investigation.
HHS AI tools and the administration’s anti-fraud push
The Trump administration and Vice President JD Vances anti-fraud task force had highlighted fraud checks in recent months. The focus included Medicaid and Medicare programmes, plus student loan applications and other areas. The effort also used AI to flag possible fraud. Federal Trade Commission Chairman Andrew Ferguson mentioned this work in remarks on Fox News.
Critics said the anti-fraud campaign had often focused on Democratic states. Some said officials appeared to accuse first and verify later. Rob Weissman, co-president of the consumer rights group Public Citizen, questioned the stated aims. \”The AI is kind of beside the point when you assess what their actual objectives are, rather than what they pretend they are,\” Weissman said.
Chiarello said safeguards mattered, but stressed the AI would review public reports. Chiarello said the system was not designed to discover new facts. Instead, officials planned to use it to spot issues faster. Chiarello also said other federal departments could reuse the approach. \”It would be fairly easy for the other agencies to use our technology and jump on it,\” Chiarello said.
With inputs from PTI
