AI task force to help guide the tech’s use in criminal justice

Olena Ruban via Getty Images
The Council on Criminal Justice and RAND Corporation are partnering to create a panel of experts that will produce research and other resources to help agencies and policymakers leverage AI safely.
The use of artificial intelligence is emerging in a range of public services, from permitting to traffic management. The tech has also made it to some court systems across the U.S. as a tool to sift through evidence, draft reports and aid other tasks to reduce workloads.
Launched earlier this week, a new task force created by the Council on Criminal Justice and research organization RAND Corporation looks to support the responsible adoption of AI products in criminal justice settings, CCJ announced Monday.
“AI is being used in law firms and in some courts to digest large volumes of material like records, depositions and transcripts … but you have to be careful,” said Nathan Hecht, chair of the task force and former Texas Supreme Court Chief Justice. “Right now, AI makes mistakes [like citing] cases that don’t exist.”
In Minnesota late last year, for instance, a legal expert’s use of generative AI to draft a court document for the state’s Attorney General’s Office garnered attention when opposing lawyers discovered the document contained AI hallucinations in the form of references to fake citations.
For criminal justice proceedings, it’s especially important to ensure any AI tools used in the process comply with people’s rights and “do not shortcut procedures that ensure justice,” Hecht said, amid the sector’s growing appetite and curiosity about the technology.
A bill proposed in Virginia this year, for instance, would prohibit AI and its content from being the “sole basis for any decision related to pre-trial detention or release, prosecution, adjudication, sentencing, probation, parole, correctional supervision, or rehabilitation of criminal offenders.”
To ensure a comprehensive review of AI’s potential and risks, the task force will also include more than a dozen AI developers, tech researchers, criminal justice practitioners, police executives, civil rights advocates and community leaders. It will also bring on people who were formerly incarcerated to help inform the task force’s efforts, its director, Jesse Rothman, said.
“If we want to really develop content with solutions and recommendations, we need to make sure that all the relevant voices are heard and represented,” he said. That means involving “people who have experience with the justice system from a range of perspectives, including formerly incarcerated, who are victims and survivors.”
The members’ work, supported by the research organization RAND Corporation, aims to develop evidence-based principles and recommendations for law enforcement agencies, courts, corrections departments and community organizations to leverage AI safely, effectively and ethically.
Those standards will inform, for instance, the task force’s efforts to develop resources aimed at improving AI procurement among criminal justice departments, policymakers and other stakeholders.
“Procurement is, in some ways, where the rubber meets the road, especially as a statutory regulatory framework,” Rothman said of organizations looking to explore how AI solutions could augment the work of court and legal systems. But it can be “quite overwhelming or uncertain about how to … understand what’s real or what’s not reliable,” he added.
With the quick pace at which AI tools and use cases change, the task force will work to research and offer procurement guidance for criminal justice departments to better navigate the market and “help guide the evolution of these departments in their thinking about these issues,” Rothman said.
Over the next 18 months, the task force will also be publishing research for policymakers, agencies and other key organizations to guide the integration of AI into current systems and processes, according to the announcement.
“AI is here — it’s all around us. It’s impacting the legal system already, and it could have an enormous impact on the criminal justice system,” Hecht said. “We just want to get it right. We want to make sure that [AI] is beneficial.”