How Computers Took Over Social Services
Virginia Eubanks experienced firsthand how automated systems can suddenly derail a person's life. After her partner, Jason, survived a brutal physical assault that required extensive facial reconstruction, their health insurance was abruptly canceled. Despite having valid coverage, an insurance company computer system claimed they had no start date, leading to the denial of over $62,000 in medical bills. While the company blamed a technical glitch, the timing and nature of the denial suggested something more calculated. Because the claims were high, filed shortly after a new policy began, and involved controlled substances for pain management, it is likely an algorithm flagged the family for a fraud investigation. This digital red-flagging happened without notice, leaving the family to battle a faceless system while trying to recover from physical trauma.
This experience highlights a massive shift in how society functions. Decisions that used to be made by people, such as who gets a mortgage, who receives health care, or who is investigated for a crime, are now handled by sophisticated machines. These automated eligibility systems and predictive risk models are often invisible and inscrutable to the general public. While they are marketed as tools for efficiency, they frequently create catastrophic hurdles for those they target. For the professional middle class, these hurdles can often be cleared with enough time, money, and persistence. However, for marginalized groups, these digital tools often create a feedback loop of injustice that reinforces their social exclusion.
The burden of this digital scrutiny falls heaviest on the poor and working class. In Maine, for example, government officials used electronic benefit records to track where low-income families withdrew cash. By highlighting a tiny fraction of transactions at liquor stores or out-of-state ATMs, politicians created a narrative of widespread fraud to justify punitive new laws. This use of data was not intended to improve the lives of the poor, but rather to stigmatize them and reinforce the idea that those receiving public assistance are inherently untrustworthy.
Across the United States, this new digital infrastructure is being integrated into public services at a rapid pace. In Indiana, welfare eligibility was automated to reduce costs. In Los Angeles, an electronic registry was created to manage the unhoused population. In Pennsylvania, predictive models now attempt to flag children at risk of abuse. These systems often discourage people from seeking the help they need and collect deeply personal information with very few privacy safeguards. They transform social problems into engineering problems, removing human empathy and discretion from the process of helping those in need.
This evolution has created what Eubanks calls the digital poorhouse. In the past, the poor were confined to physical buildings or subjected to invasive home visits by caseworkers. Today, they are managed through databases, algorithms, and risk models that track their every move. These technologies provide the public with a sense of distance, making it easier to ignore the human consequences of denying food, housing, or medical care. While these systems are currently tested on the most vulnerable populations, they establish a precedent for surveillance and automated control that could eventually be applied to everyone.
Full summary available in the Readsome app
