AI is being deployed across courts, police departments, and corrections systems. Without the right guardrails, it could amplify existing biases. But, with care and attention, there are opportunities to improve the experience of people within these same systems. Host Stephen Goldsmith speaks with Dr. Andrea Headley from Georgetown University's Evidence for Justice Lab about what governments need to know about AI in criminal justice, how to identify and reduce bias, why transparency matters for public trust, and the devastating consequences when humans aren't in the loop.