Representative Maxwell Frost (D-FL) spearheaded a letter to Governor Ron DeSantis (R-FL) and Florida Deputy Secretary for Children and Families (DCF) Kate Williams about the planned use of artificial intelligence (AI) to detect fraud in the Sunshine State's Supplemental Nutrition Assistance Program (SNAP).
The Context
In late May, it was reported that the Florida Legislature was considering the allocation of $4 million toward the implementation of AI to audit SNAP applications for eligibility and prevention of fraud.
The Problem
Rep. Frost's letter, cosigned by Representatives Kathy Castor (D-FL), Lois Frankel (D-FL), Jared Moskowitz (D-FL), Debbie Wasserman Schultz (D-FL), Darren Soto (D-FL), and Frederica Wilson (D-FL), echoes concerns other critics have raised: how will we know the AI will act fairly and accurately, and with human oversight?
Moreover, how can this be done by September 1st, which is the intended deadline for the AI's implementation?
The One Big Beautiful Bill Act (OBBBA) requires states with a SNAP error rate of over 10% to pay back 15% of the cost of benefits starting October 2027, and with Florida having a nearly 13% error rate, that means the Sunshine State could lose up to $1 billion in federal SNAP funding, potentially putting Floridians who use SNAP at risk of food insecurity.
"We understand that the state must act urgently to address the impending budgetary crisis created by the One Big Beautiful Bill Act," Rep. Frost and the others wrote. "However, we are concerned with one of the state's proposed solutions: the appropriation of $4 million in state funds for the new and relatively unproven technology of AI to identify the SNAP error rate."
They emphasized that "the state is moving on an extraordinarily accelerated timeline to procure an AI vendor by September 1, yet Floridians have received little information about how that vendor will be selected or what safeguards will be in place."
"We are seeking answers about the vendor's qualifications, the accuracy of its AI tools, protections against bias, human oversight of AI decision-making, transparency, appeals for affected SNAP recipients, and whether the state has contingency plans if it fails to meet its procurement deadline," the lawmakers concluded.
