Article Summary –
Michiganders applying for SNAP may have their cases reviewed by AI technology, intended to expedite the process and reduce payment errors, though this transition raises concerns due to Michigan’s past issues with technological systems falsely accusing individuals of fraud, such as the MiDAS system. Michele Gilman emphasizes that most payment errors are not fraudulent but often result from innocent mistakes due to the complexity of the programs, highlighting the need for careful use of the term “fraud” and transparency from government agencies. A 2024 nonprofit report warns of AI’s potential biases from those training the algorithms, potentially leading to inequities in determining eligibility for SNAP benefits.
Residents of Michigan applying for the Supplemental Nutrition Assistance Program (SNAP) might soon see their applications processed by artificial intelligence (AI). This new AI technology aims to expedite reviews and spot potential fraud, particularly over- or underpayments. On March 17, a state health official informed the Senate Appropriations Subcommittee about this development, as reported by Michigan Advance.
David Knezek, COO of the Michigan Department of Health and Human Services, explained that the AI tool would scan cases before disbursement to minimize payment error rates. In July 2025, over 1.4 million Michiganders were enrolled in SNAP, according to the Michigan League for Public Policy.
Law professor Michele Gilman from the University of Baltimore emphasized that not all payment errors are fraudulent. “Fraud involves intentional deceit for benefits, unlike most overpayments,” she stated. Many errors arise from complex SNAP procedures, involving both claimants and caseworkers.
President Donald Trump’s 2025 budget law revised SNAP by enforcing stricter eligibility, work rules, and making states cover more program costs. States with over 6% payment error rates must now contribute financially.
Gilman highlighted Michigan’s challenges with technology, referencing the Michigan Integrated Data Automated System (MiDAS). This system falsely accused 40,000 residents of unemployment fraud, leading to severe financial distress and a 400% penalty for alleged fraud. Legal actions are still unfolding despite past settlements.
As AI is increasingly used in public services, Gilman stressed the need for government transparency. Citizens deserve clear explanations regarding benefit decisions. Accountability lies with agency officials, not AI vendors.
A 2024 report from the Food Action & Resource Center warned about AI biases in SNAP eligibility determinations, caused by potentially flawed training data that could perpetuate inequities.
—
Read More Michigan News
