How AI Is Learning Bias from Predictive Policing Data

AI Ethics with Fexingo: Bias, Safety, and Responsible Artificial Intelligence · Jul 15, 2026 · 7m

In this episode of AI Ethics with Fexingo, Lucas and Luna dive into the hidden biases embedded in predictive policing algorithms. They focus on the controversial case of the PredPol system, which was trained on decades of historical crime data from the Los Angeles Police Department. The hosts break down how such data reflects over-policing in minority neighborhoods, creating a feedback loop that reinforces racial disparities. Lucas explains the concept of 'biased ground truth' — when the data used to train AI itself contains systemic bias. Luna highlights a 2023 study showing that PredPol's…