Artificial intelligence is both powerful and potentially ungovernable. It’s also already in use in criminal legal systems across the country.
While seeing significant promise for the use of AI inside legal systems—everything from data analysis to advance better policy to helping case managers connect people to services—this policy brief calls for a moratorium on any use of AI that would affect people’s liberty interests or pose a substantial risk of harm.
This brief is the second in our series on AI and the legal system, with our first publication focusing on values: what do jurisdictions want to use AI for? This newest brief follows a working session we hosted in December 2024 featuring leaders from across the technology and criminal legal spaces.
Recognizing both AI’s potential and the reality that its use will spread across local justice systems, this brief makes three principal recommendations: AI should be used only in the lowest-risk environments and decision points; any deployment must be preceded by comprehensive evaluations and simulations; and field-wide standards must be put in place to safeguard AI implementation in the criminal legal space.