As much as smartphones may engender social isolation for some, one smartphone application could help resocialize a significant U.S. population segment in need of lifelines: discharged inmates expected to reenter the social mainstream.
The National Institute of Justice (NIJ), in a 2019 grant program, is engaging researchers to find new pathways for using smartphones and other mobile devices to help offenders returning to the community. Initially, a research team from Purdue University plans to develop devices that deploy artificial intelligence (AI) to provide early warning of risky offender behavior, as well as tools to curb that behavior and help offenders comply with reentry conditions.
With AI, the intelligent machine learns and adapts as it goes. Mobile devices driven by AI, processing large data inputs, can finely tailor reentry programming to the individual. The idea behind the reentry research is that AI-equipped devices can help released inmates communicate seamlessly with their probation or parole supervisors, avoid or correct missteps, and efficiently tap into community resources best suited to their individual circumstances.
The new NIJ-sponsored study is a researcher-practitioner partnership between Purdue and Tippecanoe County, Indiana, with three main objectives:
- To develop an AI-based system to monitor and support offender reentry.
- To deploy that system with Tippecanoe County Community Corrections.
- To analyze data received from the AI-based system in order to improve the system.
The study plan calls for testing an AI-based Support and Monitoring System (AI-SMS) on offenders who will be issued smartphones and wearable biometric devices or bracelets. A total of 250 randomly selected offenders will participate in the study, with half assigned to a treatment group and the other half assigned to a control group. Selected offenders must meet a number of conditions, e.g., having been released from confinement in the last 30 days; having been assessed as at moderate to high risk of reoffending; and being willing to use the smartphone and wearable device. The impact of the AI-SMS devices will be assessed by measuring the experience of the offenders who receive them in comparison with the experience of control group members who do not receive them as part of a randomized controlled trial.
A reentry smartphone not only could monitor where an offender is going — and issue alerts to the supervision office when, for example, an offender is at a location associated with a high-risk individual or in a crime hot spot — it also could constantly accumulate data on other reentry success variables such as:
- Technical violation of reentry terms
- Re-arrest for new crimes
- Quality of participant interaction with the supervision office
- Participant goal attainment
- Community stability (e.g., stable housing and employment, connection to treatment and services support, involvement in positive social networks)
At the same time, the Purdue team plans to develop an AI-equipped biometric tracking device to monitor an offender’s health and physiological status. Such a device might monitor, for example, blood pressure, heart rate, cortisol (as a measure of stress), and body temperature.
NIJ’s Vision of Using AI to Strengthen a Wide Array of Reentry Support Services
NIJ views the Purdue grant as the first phase of research and development applying the vast power of AI to the objectives of community reentry. Offenders reentering the community through probation or parole face challenges known to elevate their risk of future incarceration. Among those challenges are unemployment, lack of financial support, substance abuse, and exposure to criminal influences. Challenges vary by state, county, city, and neighborhood. Moreover, each offender has distinguishing personal traits, tools, and experiences that make the rehabilitation process unique. One-size-fits-all thinking is a poor fit where good outcomes demand individualized approaches to reentry.
Responding to a multifaceted need, an interdisciplinary team of NIJ scientists — experts in criminology, computer science, neuroscience, and behavioral science — developed a reentry research agenda. The agenda seeks to harness the power of AI for the analysis and sharing of data generated by, and in support of, community supervision. As broadly envisioned by NIJ, new reentry hardware and software would deliver separate but symbiotic resource streams for community supervision officers and returning offenders. Proposed new technology would modernize and update the supervision capabilities of community corrections officers while focusing social networking, media, and communication apps on improving the experience of returning offenders. Research teams would articulate smartphone-based plans to enable offenders, working with their supervision officers, to construct rehabilitation strategies shaped by their own specific needs and incorporating employment, education, and treatment plans.
Smart systems can create individualized digital profiles for each offender that integrate location data, rehabilitative benchmarks, and risk and needs assessments. A system of alerts could be generated automatically, allowing the supervising officer to prioritize higher-risk individuals while giving low-risk offenders the opportunity to build confidence, discipline, and self-sufficiency. An online dashboard could give low-risk offenders a clear picture of their risks and needs, as well as access to opportunities and resources, without requiring frequent direct contact from a supervision officer.
Empowering Released Offenders to Take Control of Their Lives
One purpose of the reentry smartphone line of research is to empower offenders to take control of their lives by giving them continuous access to the information and resources they require to do so. The greater the sense of personal agency a returning offender has with respect to a rehabilitation plan, the greater the likelihood of success, studies have shown. Behavioral change is hard, but people are more likely to have success when they can see their progress and receive encouraging messages, as with many fitness trackers.
As envisioned by the NIJ scientific team behind the reentry smartphone research plan, returning offenders would gain agency as they monitor their own progress directly from their mobile devices, receive incentives and bonuses for achieving behavioral and rehabilitative benchmarks, and access assistance from a network of peer mentors, officers, or service providers in the community when they find themselves in high-risk situations.
Smartphone-facilitated reentry can also benefit supervisors responsible for assessing the risks and needs of every offender in their portfolio. Growing supervisor caseloads tend to reduce the time and attention that a single officer can devote to each assigned offender, making recidivism more likely. AI-equipped smartphones can automate and streamline the supervision workflow, using data to better predict recidivism risks and rehabilitative needs.
NIJ’s reentry smartphone research plan represents a vision for using “big data” to empower returning offenders through direct, accurate, and immediate feedback that can further their individual rehabilitative goals. At the end of the day, the goals of returning offenders and their supervision officers are generally the same: successful reentry into society and desistance from crime.
About This Article
The content of this article relates to the National Institute of Justice’s fiscal year 2019 solicitation “Artificial Intelligence Research and Development to Support Community Supervision solicitation,” requesting proposals for innovative, investigator-initiated technology research and development projects to apply advances in artificial intelligence to promote the successful reentry of offenders under community supervision; and to a grant to Purdue University pursuant to that NIJ solicitation, 2019-75-CX-K001, “AI-Enabled Community Supervision for Criminal Justice Services.”
[note 1] Matthew R. Durose, Alexia D. Cooper, and Howard N. Snyder, Recidivism of Prisoners Released in 30 States in 2005: Patterns from 2005 to 2010, Special Report, Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics, April 2014, NCJ 244205.
[note 2] Christopher T. Lowenkamp and Kristin Bechtel, “The Predictive Validity of the LSI-R on a Sample of Offenders Drawn From the Records of the Iowa Department of Corrections Data Measurement System,” Federal Probation: A Journal of Correctional Philosophy and Practice 71 no. 3 (2007).
[note 3] National Institute of Justice funding opportunity, “Artificial Intelligence Research and Development to Support Community Supervision,” grants.gov announcement number NIJ-2019-15287, posted February 15, 2019.
[note 4] Thomas H. Cohen, Christopher T. Lowenkamp, and Scott W. VanBenschoten, “Does Change in Risk Matter? Examining Whether Changes in Offender Risk Characteristics Influence Recidivism Outcomes,” Criminology & Public Policy 15 no. 2 (2016): 263-296, doi:10.1111/1745-9133.12190.
[note 5] Yilma Woldgabreal, Andrew Day, and Tony Ward, “Linking Positive Psychology to Offender Supervision Outcomes: The Mediating Role of Psychological Flexibility, General Self-Efficacy, Optimism, and Hope,” Criminal Justice and Behavior 43 no. 6 (2016): 697-721, doi:10.1177/0093854815620816.
[note 6] Patrick J. Kennealy, Jennifer L. Skeem, Sarah M. Manchak, and Jennifer Eno Louden, “Firm, Fair, and Caring Officer-Offender Relationships Protect Against Supervision Failure,” Law and Human Behavior 36 no. 6 (2012): 496-505, doi:10.1037/h0093935.