Automated Welfare Systems Wrongfully Cut Benefits in Multiple US States

Beginning in 2007, multiple US states deployed automated eligibility determination systems intended to streamline welfare administration and reduce costs. Indiana's $1.37 billion contract with IBM to modernize its Family and Social Services Administration represented the largest such initiative. The system was designed to automate decisions about Medicaid, food stamps, and Temporary Assistance for Needy Families using algorithmic processing of applications and periodic reviews. Within months of implementation, the automated systems began generating massive numbers of benefit terminations. In Indiana alone, the denial rate for food stamp applications jumped from 4% to 54% after automation. The algorithms systematically flagged legitimate beneficiaries for termination due to factors like missed phone appointments, incomplete paperwork, or failure to respond within narrow timeframes. The systems lacked contextual understanding of individual circumstances such as disabilities, language barriers, or lack of transportation that prevented compliance with rigid requirements. Arkansas implemented similar automated systems for Medicaid eligibility that used algorithmic scoring to determine benefit levels. A federal lawsuit revealed that the algorithm contained unexplained scoring rules that systematically reduced benefits for certain populations. The system assigned numerical scores based on questionnaire responses but provided no transparency into how scores translated to benefit levels. Disabled individuals and elderly patients saw dramatic cuts in home healthcare services without clear justification. The human cost was severe and immediate. Diabetics lost access to insulin, disabled individuals lost home healthcare services, and families lost food assistance. In Indiana, the percentage of applications processed within federal time requirements dropped from 90% to 27%. Beneficiaries faced months-long appeals processes to restore wrongfully terminated benefits, during which many suffered health emergencies, evictions, and other crises. Documentation from legal challenges revealed that caseworkers were pressured to meet quotas that incentivized benefit denials. Legal challenges mounted across affected states throughout the 2010s. The Indiana case resulted in a $40 million settlement in 2012 after a class action lawsuit documented systematic violations of due process. Arkansas faced a successful federal court challenge in 2018 that required the state to provide manual review of all algorithmic benefit determinations. Similar lawsuits in Idaho and Oregon led to policy changes requiring human oversight of automated decisions. The Centers for Medicare & Medicaid Services eventually issued guidance requiring states to ensure automated systems comply with federal due process requirements.