“We’ll continue to create systems that not only automate inequality but amplify inequality.”
Political scientist Virginia Eubanks has documented automated systems used in public benefits, homelessness services, and child welfare. Her warning describes more than biased code. When institutions already treat people unequally, automation can accelerate decisions, connect more data, and extend surveillance without correcting the underlying priorities. A faster process may simply deny assistance faster. A more comprehensive database may expose vulnerable families to greater scrutiny. Technology magnifies the goals and assumptions of the organization that adopts it.
People with limited resources often encounter these systems first and with the least power to refuse. They may be asked to disclose extensive personal information to receive food, housing, health care, or other necessities. Errors can trigger lost benefits, investigations, or long delays, while appeal processes remain difficult to navigate. Wealthier people are more likely to experience automation as convenience; poor and working-class communities may experience it as eligibility screening, fraud detection, or risk classification. The difference reveals how technical systems distribute both service and suspicion.
Amplification occurs through feedback. If a neighborhood receives more enforcement, it produces more records that justify additional enforcement. If an algorithm prioritizes cases resembling previous interventions, the institution sees more of the patterns it already expects. Cost-cutting can worsen the loop by reducing the staff available to understand unusual circumstances or correct data. The system then appears consistent because it repeats the same rule, even when consistency locks in an unfair starting point. Scale turns an individual mistake into a durable administrative structure.
Eubanks’s research suggests several safeguards: include affected communities before procurement, minimize unnecessary data, publish understandable rules, maintain accessible human review, and measure whether people actually receive needed help. Officials should compare automation with investments in staff, services, and simpler eligibility policies. Sometimes the better solution is not a smarter screening model but fewer barriers and more resources. Accountability must also remain public. A vendor’s proprietary system cannot become an excuse for a government agency to withhold reasons from the people it serves. Eubanks’s warning is not that every automated tool must fail. It is that technology inherits the purpose and power of the institution using it. Reform must address both the system and the conditions the system is built to manage. A humane interface cannot compensate for inadequate benefits, punitive eligibility rules, or a lack of affordable housing. Technology should be evaluated against those structural alternatives rather than accepted as the only available form of modernization.
Software inherits the goals, constraints, and histories of the agency or organization that deploys it.
Speed and scale can intensify an unjust process even when the code executes its specification accurately.
People seeking essential services often cannot refuse data collection, wait through errors, or hire help for an appeal.
The same technology experienced as convenience by one group may function as surveillance and exclusion for another.
Public agencies should evaluate staffing, simpler rules, direct services, and reduced data collection alongside automated tools.
A smarter screening system is not automatically better than removing unnecessary barriers to assistance.
Explore more "Quotes of The Day"
Continue exploring how writers, thinkers, artists, and advocates have described the choices that shape an everyday life.