Criminal justice agencies should articulate clear, values-driven goals when considering the adoption of artificial intelligence tools, the think tank Council on Criminal Justice (CCJ) urged in a report issued on Tuesday. These may include increasing efficiency by automating routine tasks to free up staff time for higher-value work or improving equity by using AI to identify and mitigate human biases or errors in decision-making. Agencies should be explicit about the intended purposes upfront and assess any potential AI application against them, CJJ said. The report was based on a national convening of researchers, policymakers, advocates, and other stakeholders, hosted by CCJ and the Stanford Criminal Justice Center. Participants discussed the potential of AI to improve criminal justice outcomes while warning of the complex challenges and risks associated with AI tools.
One recommendation in the report was that criminal justice agencies engage with AI systems, including private companies that are producing AI tools and technologies, to shape their development and implementation in ways that align with public safety goals and ethical considerations. The report also urged providing the necessary training and preparation for the thousands of criminal justice leaders and independent agencies to develop a solid grasp of AI fundamentals to make informed decisions about adopting and governing these tools. "Criminal justice agencies should recognize that AI systems will almost certainly become more sophisticated and common over time," CJJ said. "Rather than adopting a reactive posture, agencies should proactively engage with AI technologies.". Core elements for criminal justice AI governance, CJJ said, could include conducting impact assessments for proposed AI applications, applying more stringent controls to higher-risk use cases, ensuring transparency and preserving meaningful human oversight, and building in ongoing monitoring and evaluation.
Comments