Black box machine learning models are currently being used for high-stakes decision making throughout society, causing problems in healthcare, criminal justice and other domains. Some people hope that creating methods for explaining these black box models will alleviate some of the problems, but trying to explain black box models, rather than creating models that are interpretable in the first place, is likely to perpetuate bad practice and can potentially cause great harm to society. The way forward is to design models that are inherently interpretable. This Perspective clarifies the chasm between explaining black boxes and using inherently interpretable models, outlines several key reasons why explainable black boxes should be avoided in high-stakes decisions, identifies challenges to interpretable machine learning, and provides several example applications where interpretable models could potentially replace black box models in criminal justice, healthcare and computer vision. There has been a recent rise of interest in developing methods for ‘explainable AI’, where models are created to explain how a first ‘black box’ machine learning model arrives at a specific decision. It can be argued that instead efforts should be directed at building inherently interpretable models in the first place, in particular where they are applied in applications that directly affect human lives, such as in healthcare and criminal justice.

Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead

Nos últimos trinta trinta anos, houve profundas mudanças na forma como compreendemos o crime e a justiça criminal. O crime tornou-se um evento simbólico, um verdadeiro teste para a ordem social e para as políticas governamentais, um desafio para a sociedade civil, para a democracia e para os direitos humanos. Segundo David Garland, professor da Faculdade de Direito da New York University, um dos principais autores no campo da Sociologia da Punição e com artigo publicado na Revista de Sociologia e Política , número 13, na modernidade tardia houve uma verdadeira obsessão securitária, direcionando as políticas criminais para um maior rigor em relação às penas e maior intolerância com o criminoso. Há trinta anos, nos EUA e na Inglaterra essa tendência era insuspeita. O livro mostra que os dois países compartilham intrigantes similaridades em suas práticas criminais, a despeito da divisão racial, das desigualdades econômicas e da letalidade violenta que marcam fortemente o cenário americano. Segundo David Garland, encontram-se nos dois países os “mesmos tipos de riscos e inseguranças, a mesma percepção a respeito dos problemas de um controle social não-efetivo, as mesmas críticas da justiça criminal tradicional, e as mesmas ansiedades recorrentes sobre mudança e ordem sociais”1 (GARLAND, 2001, p. 2). O argumento principal da obra é o seguinte: a modernidade tardia, esse distintivo padrão de relações sociais, econômicas e culturais, trouxe consigo um conjunto de riscos, inseguranças e problemas de controle social que deram uma configuração específica às nossas respostas ao crime, ao garantir os altos custos das políticas criminais, o grau máximo de duração das penas e a excessivas taxas de encarceramento.

http://www.scielo.br/pdf/rsocp/n20/n20a15.pdf

The culture of control: crime and social order in contemporary, society

Machine learning is a means to derive artificial intelligence by discovering patterns in existing data. Here, we show that applying machine learning to ordinary human language results in human-like semantic biases. We replicated a spectrum of known biases, as measured by the Implicit Association Test, using a widely used, purely statistical machine-learning model trained on a standard corpus of text from the World Wide Web. Our results indicate that text corpora contain recoverable and accurate imprints of our historic biases, whether morally neutral as toward insects or flowers, problematic as toward race or gender, or even simply veridical, reflecting the status quo distribution of gender with respect to careers or first names. Our methods hold promise for identifying and addressing sources of bias in culture, including technology.

/pdf/semantics-derived-automatically-from-language-corpora-4lzxrze2vb.pdf

Semantics derived automatically from language corpora contain human-like biases

With the widespread use of AI systems and applications in our everyday lives, it is important to take fairness issues into consideration while designing and engineering these types of systems. Such systems can be used in many sensitive environments to make important and life-changing decisions; thus, it is crucial to ensure that the decisions do not reflect discriminatory behavior toward certain groups or populations. We have recently seen work in machine learning, natural language processing, and deep learning that addresses such challenges in different subdomains. With the commercialization of these systems, researchers are becoming aware of the biases that these applications can contain and have attempted to address them. In this survey we investigated different real-world applications that have shown biases in various ways, and we listed different sources of biases that can affect AI applications. We then created a taxonomy for fairness definitions that machine learning researchers have defined in order to avoid the existing bias in AI systems. In addition to that, we examined different domains and subdomains in AI showing what researchers have observed with regard to unfair outcomes in the state-of-the-art methods and how they have tried to address them. There are still many future directions and solutions that can be taken to mitigate the problem of bias in AI systems. We are hoping that this survey will motivate researchers to tackle these issues in the near future by observing existing work in their respective fields.

A Survey on Bias and Fairness in Machine Learning

Recidivism prediction instruments (RPIs) provide decision-makers with an assessment of the likelihood that a criminal defendant will reoffend at a future point in time. Although such instruments are gaining increasing popularity across the country, their use is attracting tremendous controversy. Much of the controversy concerns potential discriminatory bias in the risk assessments that are produced. This article discusses several fairness criteria that have recently been applied to assess the fairness of RPIs. We demonstrate that the criteria cannot all be simultaneously satisfied when recidivism prevalence differs across groups. We then show how disparate impact can arise when an RPI fails to satisfy the criterion of error rate balance.

Fair Prediction with Disparate Impact: A Study of Bias in Recidivism Prediction Instruments

COMMUNITY SUPERVISION IS one of the most widely imposed court responses, with approximately 5,095,200 or 70 percent of the correctional population being under community supervision (Glaze, 2010). Despite its popularity, researchers have limited insight into whether community supervision is an effective strategy for reducing recidivism. The most recent reviews of the effectiveness of community supervision (Solomon, Kachnowski, & Bhati, 2005; Aos, Miller, & Drake, 2006; Bonta, Rugge, Scott, Bourgon, & Yessine, 2008; Green & Wink, 2010) brought sobering results. More than three decades after Martinson (1974) summarized the findings of his review of rehabilitation efforts by saying \"with few and isolated exceptions, the rehabilitative efforts that have been reported so far have had no appreciable effect on recidivism,\" Bonta and his colleagues (2008) have found that more recent research yields no better results: the impact of community supervision is limited at best and non-existent in the most pessimistic interpretation.

A random (almost) study of staff training aimed at reducing re-arrest (STARR): Reducing recidivism through intentional design.

Level of risk is proving to be an important characteristic in effectively serving offender populations. A major limitation to the use of risk assessments is agency resources. There are several screening instruments available that could significantly decrease the amount of resources that are needed to assess for risk. This article assesses the effectiveness of the Level of Service Inventory: Screening Version on 483 probationers in a western state resulting in a 2% false-positive rate. Policy implications are explored and suggestions for future research offered.

Validating the Level of Service Inventory—Revised and the Level of Service Inventory: Screening Version With a Sample of Probationers:

Offender assessment has been and remains the cornerstone of effective community supervision. This article presents the development of and tests the predictive validity of a 4th-generation risk assessment instrument designed for U.S. probation. A large administrative data set was used to create the assessment instrument and conduct an initial validation. Subsequent data generated from officer-completed assessments were used to conduct a prospective validation. Finally, data from case vignettes scored by trained officers were used to test the interrater agreement of the assessment instrument. Overall, analysis revealed that the assessment instrument predicted rearrest reliably when using the assessment results based on administrative data or officer-completed assessments. Analysis also revealed high rates of interrater agreement. Recommendations for future research and policy implications are presented.

The federal Post Conviction Risk Assessment (PCRA): a construction and validation study.

Although prior research suggests that juvenile correctional programs can be effective at reducing recidivism, research has also found that some programs are more effective than others. This suggests that identifying the characteristics of effective correctional programs is an important issue surrounding interventions with juveniles. The current research not only examines the effectiveness of Ohio’s community correctional facilities in reducing recidivism, but it also examines whether measures of treatment integrity are related to recidivism. The findings revealed that programs that scored higher on treatment integrity were more effective at reducing recidivism. Specifically, community corrections facilities that served higher risk youth, targeted dynamic risk factors with cognitive behavioral modalities, and employed trained and qualified staff were found to have stronger effects on recidivism than those that did not.

https://www.uc.edu/content/dam/uc/ccjr/docs/articles/CCF_Juvenile.pdf

Community Corrections Facilities for Juvenile Offenders in Ohio An Examination of Treatment Integrity and Recidivism

In Rethinking Rehabilitation, Farabee claims that offender treatment is a failed enterprise and instead proposes a correctional approach that emphasizes deterrence through intensive supervision, electronic monitoring, and indeterminate parole sentences. We argue that this neo-Martinson attack on rehabilitation, which has the potential to shape public policy discourse, needs to be deconstructed. Although Farabee's critique has merits—especially about the limited effectiveness of many current prison programs—his analysis ignores research both favorable to offender treatment and unfavorable to his proposed policy agenda. In this context, his advice to choose a correctional future that is punitive and devoid of rehabilitation would be a mistake.

Christopher T. Lowenkamp

Papers

A random (almost) study of staff training aimed at reducing re-arrest (STARR): Reducing recidivism through intentional design.

Validating the Level of Service Inventory—Revised and the Level of Service Inventory: Screening Version With a Sample of Probationers:

The federal Post Conviction Risk Assessment (PCRA): a construction and validation study.

Community Corrections Facilities for Juvenile Offenders in Ohio An Examination of Treatment Integrity and Recidivism

Nothing Works Revisited: Deconstructing Farabee's Rethinking Rehabilitation