Experts Warn AI Vetting of Michigan SNAP Applications May Generate False Fraud Claims

Michigan’s health department is deploying artificial intelligence to review applications for the state’s Supplemental Nutrition Assistance Program (SNAP), a move officials claim will accelerate processing times and identify potential fraud cases.

David Knezek, chief operating officer of the Michigan Department of Health and Human Services, unveiled the AI initiative during a state Senate Appropriations Subcommittee meeting on March 17. Knezek explained the technology would allow comprehensive scanning of all applications before benefits are disbursed.

“Using this AI case reading tool, we’re now not only going to be able to scan every single case in a perfect environment before that money goes out the door, we’re also going to be able to target it to the cases that have the highest likelihood of resulting in a payment error rate,” Knezek told lawmakers.

The program affects a substantial number of Michigan residents. According to the Michigan League for Public Policy, more than 1.4 million Michiganders participated in SNAP as of July 2025.

Legal experts, however, have raised concerns about the implementation of AI for benefit determination. Michele Gilman, professor of law at the University of Baltimore School of Law, emphasized to the Michigan Independent that payment errors shouldn’t automatically be classified as fraud.

“Fraud is when someone is intentionally trying to claim benefits that they are not entitled to. That’s not what is going on in the vast majority of overpayments,” Gilman explained. “Overpayments and underpayments — which are also a problem, but one we don’t hear much about — are usually the result of innocent errors.”

The push for stricter screening comes amid significant changes to the federal SNAP program. President Donald Trump’s budget bill signed into law on July 4, 2025, implemented tighter eligibility and work requirements while shifting more financial responsibility to states. Under the new law, states with payment error rates exceeding 6% must shoulder a percentage of SNAP benefit costs, which were previously funded entirely by the federal government.

Michigan’s history with automated benefit systems has been problematic. Gilman pointed to the state’s Michigan Integrated Data Automated System (MiDAS), implemented in 2013 to detect unemployment fraud, as a cautionary tale.

“Michigan, in particular, has a troubled history with using technological systems to determine eligibility and fraud,” she noted.

The MiDAS system falsely accused 40,000 Michigan residents of unemployment fraud. Under Michigan law at the time, those accused faced severe penalties — repayment of all benefits plus an additional 400% penalty. The financial consequences were devastating.

“People immediately went into extreme financial distress to pay those fees back, resulted in bankruptcies, divorces, all sorts of crisis situations,” Gilman said. A subsequent state audit found that at least 93% of the fraud accusations were unfounded, leading to class-action lawsuits that have continued for over a decade.

As government agencies increasingly adopt AI systems to streamline operations, Gilman stressed the importance of transparency and accountability.

“When you are a citizen dealing with the government over resources you need to survive, you are constitutionally entitled to a clear explanation that you can understand as to why your benefits are granted, cut, or denied,” she said. “At the end of the day, it’s the agency officials who are accountable and must be accountable for these systems. They cannot pass the buck to the vendors who design them.”

A 2024 report from the nonprofit Food Action & Resource Center highlighted specific risks associated with using AI to determine SNAP eligibility. One significant concern is the potential for human bias in the machine learning process.

“There are possibilities for malicious actors, or people who are not well-versed in SNAP, to contribute to training algorithms with inaccurate, incomplete, statistically insignificant, or biased information,” the report cautioned. Such biases could perpetuate stereotypes and inequities based on race, gender, immigration status, or other identity categories, potentially preventing eligible individuals from accessing benefits.

Michigan’s implementation of this AI system represents a growing trend of algorithmic decision-making in public benefits administration, raising important questions about accuracy, fairness, and accountability in systems affecting vulnerable populations.