The rise of artificial intelligence (AI) represents both opportunities and challenges for a fair and just criminal legal system. The Wilson Center addresses these challenges through two key areas of work: ensuring proper safeguards around AI use and evaluating emerging AI technologies to determine their effectiveness and impact.
Ensuring Safeguards for AI
The Wilson Center works to advance protections around due process and as AI becomes more prevalent in criminal legal settings. Most AI systems are designed as “black boxes” that lack transparency and interpretability. When these opaque systems influence decisions about people’s lives and liberty, they undermine basic principles of fair treatment: if people are harmed by AI-driven decisions, they have no fair notice of how the decision was made, no ability to meaningfully respond, and no way to identify or correct mistakes .
The Center has conducted leading research and policy work to prevent the spread of black box AI. Professor Brandon L. Garrett, faculty director of the Wilson Center, and Professor Cynthia Rudin, Gilbert, Louis, and Edward Lehrman Distinguished Professor of Computer Science at Duke, have led efforts to evaluate fully interpretable, “glass box” AI. This work demonstrates the need for interpretable AI in the criminal system and shows how transparent AI systems can better protect fairness and constitutional rights. Based on this research, Garrett and Rudin provide policy guidance and consulting to legal practitioners and policymakers working on AI implementation.
Selected Publications
- AI and Criminal Procedure Rights: A Response to the National Institute of Justice Request for Input By: Brandon L. Garrett, Alicia Carriquiry, Karen Kafadar, Robin Mejia, Cynthia Rudin, Nicholas Scurich, and Hal Stern (2024)
- The Right to a Glass Box: Rethinking the Use of Artificial Intelligence in Criminal Justice By: Brandon Garrett and Cynthia Rudin in the Cornell Law Review (2024)
- Interpretable Algorithmic Forensics By: Brandon Garrett and Cynthia Rudin in PNAS (2023)
Evaluating AI and Other Technology
The Wilson Center conducts research evaluations of AI use in criminal legal settings and encourages local governments to evaluate these tools before full implementation. We completed a comprehensive evaluation of ShotSpotter, a gunshot detection technology, in Durham, North Carolina, and are currently conducting a similar study in Fayetteville, North Carolina. Our Durham findings are available in our published report. We found that ShotSpotter's impact on public safety remains unclear—while it improved some police response metrics, researchers could not determine whether it actually reduced
The Center partners with communities to conduct independent evaluations and field studies of AI technology in criminal legal settings. Learn more about our evaluation process and how to partner with us.