Michigan is set to deploy a major new software system to prisons across the state, meant to classify incarcerated people based on “how dangerous they are.”
That’s according to Detroit channel Fox2, which reports that the Michigan Department of Corrections is plowing ahead with an inmate categorization algorithm that will sort people into various housing levels.
In Michigan, incarcerated people are separated into five housing levels according to their security classification. In the past, screenings were performed by a committee of staff appointed by the Warden at each facility. Now, it seems the state wants to outsource that decision — or at least have it be heavily informed by — a non-profit AI and software company called Recidiviz.
Recidiviz began as a data organization project at Google, meant to “uncover how technology could standardize criminal justice data and improve the system nationwide.” Its AI-integrated software platform is said to perform features like identifying individuals eligible for state-offered programs, reclassification, or release, and coordinate access to resources like transitional housing and employment.
In a since-deleted press release, MDOC said the Recidiviz software will be used to “optimize placement of incarcerated individuals” starting with a pilot program in a number of men’s facilities beginning in December. The state agency notes that the tool is expected to “improve MDOC’s ability to more accurately predict incarcerated individuals’ risk of future violence by approximately 30 percent,” though how the state arrived at that vague figure is unclear.
In comments to Futurism, both MDOC and Recidiviz denied that the specific piece of software being workshopped for Michigan prisons — which they termed a “classification module” — uses AI to administer placement decisions.
“AI is never used in Recidiviz tools to make decisions about anyone’s security level or release,” a spokesperson for the company claimed.
That said, it’s hard to ignore the fact that AI features prominently on Recidiviz’s website and in its marketing material.
One explainer states that “the Recidiviz Corrections Platform uses data, automation, and AI to save staff time,” while an info-tab on the company’s “Intake” capabilities states that Recidiviz allows corrections officials to “administer AI-powered conversational assessments, letting residents respond at their own pace and in their preferred language while still capturing the essential data staff need.”
Yet another page on the company’s website specifically about AI use states that “AI offers a way to reduce repetitive work, surface important insights faster, and improve consistency across teams.” Further down, the company states that it uses AI for things like “generating code across our open-source platforms” and “matching people with jobs, services, or support based on their needs.”
In a separate statement, MDOC public information officer Jenni Riehle asserted that “while Recidiviz builds AI tools that support staff with tasks such as meeting transcription, AI is never used to make decisions about an individual’s security level or release.”
But without a thorough look under the module’s hood, questions remain around where exactly in its platform Recidiviz deploys AI — and there is much at stake in the state’s decision.
For incarcerated people, housing designations are extremely consequential, determining which programming and resources they’ll be eligible for, and what kind of population they’ll be housed with for the duration of their sentence.
As a ProPublica investigation into a similar risk-analysis software in Louisiana found, the kind of data being used to classify incarcerated people contained inherent racial and socioeconomic bias.
As a result, Louisiana’s carceral system has become even more cruel and harmful to those earnestly working toward rehabilitation. In one case reported by ProPublica, a blind and wheelchair bound 70-year-old was determined to be a “moderate risk” of reoffending despite having a clean disciplinary record — a classification which results in an automatic loss of eligibility for parole.
While Michigan’s system is geared toward housing assignments rather than parole decisions, it seems highly unlikely that software alone will solve the state’s ongoing corrections crisis — and if Louisiana’s embrace of techno-solutionism is any example, it could very well make matters far, far worse.
This article has been updated to include comments from MDOC and Recidiviz.
More on incarceration: What Are the Ethics of Strapping VR Headsets on Inmates in Solitary Confinement?