This first in a series of four technology briefs that explore the use of artificial intelligence (AI) in the criminal justice system defines AI and common AI terms, and provides a mental model for identifying cases of AI use in the criminal justice system.
AI is defined as “a broad discipline encompassing a wide range of methods that seek to create machines that mimic human intelligence.” This brief is not a comprehensive guide to the various domains or underlying computational methods that enable AI applications, but it does review common AI terms and their relationship to one another. This brief also poses the following question for the criminal justice community: “How can we be proactive co-creators of an intelligent criminal justice future that balances the risks and opportunities of this new technology?” Although this brief presents examples of how AI has potential to bring significant benefit to the criminal justice system, it also identifies risks that policymakers should consider when developing or deploying AI tools. AI tools have the potential to improve efficiency, reduce costs, and expand capabilities across many criminal justice domains; however, technical feasibility and operational realities must be considered; for example, AI technologies and their applications must address ethical and civil liberties issues, given their potential for the collection, storage, and misuse of data on individuals processed or contacted by a criminal justice agency. This brief presents a “design thinking” approach for evaluating cases of AI use. This is an approach to innovation that emphasizes understanding the problem, context, and end users before designing solutions. When evaluating AI’s role in the criminal justice system, a design-thinking approach can prevent the use of AI where it is not needed or beneficial. 6 figures and 27 references