A California bill replaces cash bail with risk-assessment algorithms, which critics argue will perpetuate the pre-trial detention of minority and low-income defendants.
When the underlying data they rely on is incomplete—and it often is—the growing use of machine learning tools in America's criminal justice system can have devastating effects.
A recent vote over a proposed tool to predict the risk that a person would pose a threat to public safety in Pennsylvania stirred a debate over its unintended consequences.
The criminal justice system has been using predictive algorithms for decades, but research shows even the best algorithms are no better than humans at predicting recidivism—and neither are very good.