Machine Bias

Machine Bias is a landmark investigative article examining the COMPAS (Correctional Offender Management Profiling for Alternative Sanctions) algorithm, which is used across the United States to assess the likelihood that criminal defendants will reoffend. ProPublica's analysis of over 10,000 criminal cases in Broward County, Florida revealed significant racial disparities: the algorithm was nearly twice as likely to incorrectly flag Black defendants as high-risk compared to white defendants, while overestimating risk for white defendants. The investigation combined data journalism, interviews with defendants and officials, and statistical analysis to demonstrate how ostensibly objective algorithmic decision-making can perpetuate and amplify systemic racial bias in the criminal justice system.

The article sparked widespread debate about algorithmic accountability, fairness in machine learning, and the role of automated systems in high-stakes decisions affecting human lives. It became a foundational text in discussions of AI ethics and bias, influencing academic research, policy discussions, and corporate practices around algorithmic transparency and fairness testing.