The Michigan Department of Corrections is set to test a new security classification system that aims to enhance the prediction of violent behavior among inmates. This innovative initiative, scheduled to pilot on December 1, leverages the analysis of historical prison data while foregoing traditional artificial intelligence models. Instead, the system employs a rules-based approach designed to classify incarcerated individuals based on their behavior, increasing objectivity while mitigating the subjective nature of manual assessments. As this system rolls out, it raises important questions about the accuracy of predictions and the overall safety within Michigan’s correctional facilities.
| Article Subheadings |
|---|
| 1) Why security levels can change life behind bars |
| 2) What Michigan’s new scoring system will change |
| 3) Is Michigan actually using AI to predict prison violence? |
| 4) Can Michigan really predict prison violence 30% better? |
| 5) Why better prison technology may not solve Michigan’s biggest problem |
Why security levels can change life behind bars
The Michigan Department of Corrections is responsible for managing approximately 32,000 inmates across its facilities. One of the crucial elements of maintaining order and security within these institutions lies in the classification of inmates based on their security levels. Currently, corrections officials utilize a manual scoring system to allocate security levels, which dictates not only the level of supervision an inmate receives but also their access to educational and vocational training programs. For inmates, especially those serving lengthy sentences, any changes to their security classification can significantly impact their everyday lives, including social interactions, work assignments, and rehabilitation opportunities.
In Michigan, security classifications range from Level I to Level V, with Level I allowing greater freedom of movement and access to resources, while Level V involves more stringent control measures. These classifications can influence an inmate’s daily routines and overall experience in the correctional system. Recognizing the importance of these classifications, the Michigan Department of Corrections is seeking an improved methodology to assess which security level is most appropriate for individual inmates based on their current behaviors and rehabilitation needs.
What Michigan’s new scoring system will change
The impending pilot of Michigan’s new scoring system is expected to revolutionize the way correctional staff conduct assessments of inmate security levels. Set to commence on December 1, the pilot program will take place in select men’s prisons with the goal of providing ongoing evaluations of inmates rather than relying solely on scheduled assessments. This adaptation allows for more timely recognition of behavioral changes, thus improving the safety and management of the prison environment.
Developed by Recidiviz, a nonprofit organization specializing in criminal justice technology, the new system utilizes insights from a decade’s worth of classification data collected from Michigan prisons. The software assesses inmates’ current records, allowing corrections staff to calculate security scores based on established factors. For instance, recent disciplinary issues may carry greater weight in determining an inmate’s classification compared to past infractions. Furthermore, inmates who maintain good behavior over an extended period may gain eligibility for a decreased security classification.
Moreover, the system is designed to trigger reevaluations whenever new information arises, ensuring that the evaluations remain relevant and aligned with the inmate’s behavior. Michigan Department of Corrections Director Heidi E. Washington has expressed confidence that these changes will bolster safety measures within the facilities. She remarked,
“Classifying prisoners at the security level that fits their current behaviors and needs keeps everyone safer.”
Is Michigan actually using AI to predict prison violence?
In recent discussions, there has been confusion surrounding the role of artificial intelligence in Michigan’s new security classification system. On November 2, Recidiviz clarified that the scoring process does not employ AI to generate security scores. Instead, it utilizes a points-based formula that is thoroughly examined by corrections personnel. This fact challenges prevailing narratives that suggest AI plays a central role in risk assessment within correctional facilities.
The scoring system can be thought of as a calculated scorecard; it compiles data from inmate records and assigns scores according to fixed rules established by corrections experts. This approach maintains human oversight throughout the evaluation process, allowing corrections staff to review and amend security scores when necessary. Situations involving medical needs or personal safety concerns are duly considered before reaching final decisions on a housing assignment—for instance, if an inmate’s behavior does not align with their recorded classification.
While Recidiviz implements AI technologies in other facets of its platform to assist corrections personnel with administrative tasks, it has expressly defined that the classification of inmates and their subsequent housing decisions are guided by a traditional app-based methodology. This detail is critical, as it sheds light on the relationship between technology and correctional decision-making processes, emphasizing the importance of understanding how assessments impact inmates’ lives.
Can Michigan really predict prison violence 30% better?
A claim issued by the Michigan Department of Corrections stated that the new classification system could increase the prediction accuracy of future violent incidents by approximately 30%. While this assertion appears promising, it is imperative to demand greater clarity on how this figure was derived and what specific metrics were used to determine the current performance standards.
Critical questions arise from this assertion: Will the system successfully identify more individuals who pose potential risks of violence? Conversely, could it lead to incorrect assumptions about an inmate’s current risk level? Furthermore, what methodology will officials employ to measure the effectiveness of the new system when compared to existing classification processes?
The reliability of the software is closely linked to the accuracy of the underlying prison records. If there are inaccuracies in disciplinary reports, these could negatively influence an inmate’s score. Additionally, it must be noted that no risk assessment tool can guarantee precise predictions regarding future behavior, thus underscoring the need for human intervention at all stages of the process.
Why better prison technology may not solve Michigan’s biggest problem
The anticipated improvements in security classification brought by the new technology do not come without challenges. The Michigan Department of Corrections, even with a more accurate scoring system, faces a significant issue regarding the actual housing capacity within its facilities. Concerns have been raised about a deficiency in available beds in high-security environments, which could negate the benefits of any new predictive technology.
As highlighted by Byron Osborn, the president of the union representing Michigan corrections officers, the real challenge lies in the capability of the department to match identified needs with appropriate security housing. If an assessment indicates that an inmate requires tighter supervision but the facility lacks space for them, it raises the question: what next? It points to a systemic issue that cannot solely be resolved by technological advancements.
In response to the growing necessity for more secure housing, Michigan has plans to add approximately 870 high-security beds as part of its Safe Prisons Initiative. Some facility upgrades are already underway, while other larger projects are projected to take time before completion. However, inmates have already experienced instances of violence, including fatalities, making it clear that the need for timely solutions is urgent and that technological improvements must translate into effective resource management.
| No. | Key Points |
|---|---|
| 1 | Michigan’s new security classification system focuses on improving inmate safety by enabling ongoing evaluations of security levels. |
| 2 | The scoring system is not powered by artificial intelligence but rather employs a points-based formula reviewed by corrections staff. |
| 3 | The expected prediction improvement of violent incidents is approximately 30%, prompting calls for transparency on how this figure was calculated. |
| 4 | Challenges remain regarding the availability of high-security beds in Michigan prisons, complicating the implementation of the new technology. |
| 5 | Human oversight is a critical component of the new system, ensuring that security classifications consider individual circumstances. |
Summary
The upcoming pilot of Michigan’s security classification system represents a significant shift in the approach to inmate management. While the potential for increased prediction accuracy in violent behavior is noteworthy, it raises important questions regarding the evaluation processes, the role of technology, and the fundamental challenges that still exist in the correctional environment. Enhanced safety measures will require not just better systems but also adequate facilities and resources to meet the identified needs of incarcerated individuals. As the Michigan Department of Corrections moves forward, the effectiveness of these technological advancements will be closely scrutinized.
Frequently Asked Questions
Question: How will the new security classification system affect inmates in Michigan?
The new scoring system is designed to enable ongoing evaluations, which can lead to changes in an inmate’s security level based on their current behavior and needs. This can impact their daily lives significantly, influencing their access to programs and levels of supervision.
Question: What role does technology play in the new classification system?
The system utilizes a points-based formula derived from historical data and does not rely on artificial intelligence to calculate security scores. Corrections staff will manually review and make final decisions based on the generated scores.
Question: What are the biggest challenges facing the implementation of this system?
One of the primary challenges is the shortage of available high-security beds, which limits the ability to house inmates according to their assessed needs. This gap may undermine the effectiveness of the new classification system.
