Predictive policing algorithms are racist. They need to be dismantled.
The article illustrates how data systems can perpetuate systemic racism, tracing the issue back to incidents where students were disproportionately arrested and subsequently targeted by biased algorithms. This personal history underscores a broader truth: data is often weaponized against Black communities, transforming individual encounters into lifelong markers of criminality. The narrative highlights the critical role of data activism in exposing these injustices and challenging the infrastructure that fuels them. Predictive policing tools, which analyze location trends or personal histories to forecast criminal behavior, are shown to reinforce existing inequalities. Because these algorithms rely on historical arrest data, they inherently reflect the over-policing of minority groups. Consequently, even when race is not explicitly included, proxy variables like zip codes and socioeconomic status allow these systems to discriminate effectively. This mechanism creates a feedback loop where marginalized communities are continuously flagged and monitored, validating the premise that the technology itself is fundamentally flawed. This perspective is vital for the open data community as it reveals the real-world harm caused by opaque algorithmic decision-making. It demonstrates that open data initiatives must address not just accessibility, but also equity and the potential for bias within datasets. Understanding these dynamics is essential for developing ethical data practices that dismantle, rather than deepen, the school-to-prison pipeline and other structural inequities.
Source: technologyreview.comPublished on 2023-10-11