A case study of algorithm-assisted decision making in child maltreatment hotline screening decisions

By Alexandra Chouldechova, Diana Benavides Prado, Oleksandr Fialko, Rhema Vaithianathan. 2018

339 citations recorded by the source corpus.

Subfields: AI Ethics, Fairness & Society, Social Sciences & Humanities

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In AI Ethics, Fairness & Society

In Social Sciences & Humanities

Citations in the corpus

Designing Disaggregated Evaluations of AI Systems: Choices, Considerations, and Tradeoffs (2021, methodology): “[17] used records that were previously collected by Allegheny County to assess and report different models’ area under the curve for groups based on race.”
Racial disparities in automated speech recognition (2020, background): “There is worry, however, that speech recognition systems suffer from racial bias (1–4), a problem that has recently come to light in several other advancing applications of machine learning, such as face recognition (5, 6), natural language processing (7–11), online advertising (12, 13), and risk pr”
A Survey on Bias and Fairness in Machine Learning (2019, background): “ue for researchers and engineers. Machine learning is used in courts to assess the probability that a defendant recommits a crime. It is used in different medical fields, in childhood welfare systems [32], and autonomous vehicles. All of these applications have a direct effect in our lives and can h”
Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions (2018, background): “For example, in pretrial decisions, the sample may be drawn from the population conditional on release, excluding people who were detained (a problem sometimes called selective labels) [98, 105, 22, 29, 37].”
Aequitas: A Bias and Fairness Audit Toolkit (2018, background): “Fairness has been defined has being being accountable on the treatment Pedreshi et al. (2008); Dwork et al.”
Learning Optimal Fair Policies (2018, methodology): “As a motivating example, we consider a simpliﬁed model for a children’s welfare screening program, recently discussed in (Chouldechova et al. 2018; Hurley 2018).”
A Causal Bayesian Networks Viewpoint on Fairness (2018, neutral): “…scenarios that have serious implications for individuals and society, including ﬁnancial lending [10, 44], hiring [8, 27], online advertising [26, 38], pretrial and immigration detention [5, 40], child maltreatment screening [13, 45], health care [18, 31], and social services [1, 22].”
Characterizing Fairness Over the Set of Good Models Under Selective Labels (2021, background): “Algorithmic risk assessments are used to inform decisions in high-stakes settings such as health care, child welfare, criminal justice, consumer lending and hiring [1, 2, 3, 4, 5].”
Actionable Recourse in Linear Classification (2018, background): “that it should be. A model should provide an individual with recourse when it is used to allocate a good that should be universally accessible, such as credit [29], employment [2] and public services [8, 28]. However, recourse should also be considered more broadly given that a lack of human agency ”
Improving Fairness in Machine Learning Systems: What Do Industry Practitioners Need? (2018, background): “public-sector ML practitioners working across a range of high-stakes contexts, such as predictive policing [65, 71, 94] and child mistreatment detection [24], to understand the challenges they face in aligning the behavior of ML systems”
Closing the AI accountability gap: defining an end-to-end framework for internal algorithmic auditing (2020, background): “dit framework on two hypothetical client projects. The first (hypothetical) client wishes to develop a child abuse screening tool similar to that of the real cases extensively studied and reported on [11, 14, 15, 21, 25, 36]. This complex case intersects heavily with applications in high-risk scenar”
Fairness for Unobserved Characteristics: Insights from Technological Impacts on Queer Communities (2021, background): “i and Gebru 2018), reduce gender bias in language processing (Bolukbasi et al. 2016; Park, Shin, and Fung 2018), and increase the accuracy and equity of decision making for child protective services (Chouldechova et al. 2018). 1Throughout this paper, we distinguish between ‘legal gender’ (the gender”
Residual Unfairness in Fair Machine Learning from Prejudiced Data (2018, background): “(3) Risk assessment and intervention tools in child welfare agencies are trained on cases which have been “screened in” by caseworkers based on external reports (Chouldechova et al., 2018).”
Simplicity Creates Inequity: Implications for Fairness, Stereotypes, and Interpretability (2018, neutral): “We are already seeing algorithms play this role in domains including hiring, education, lending, medicine, and criminal justice [8, 22, 34, 36].”
Active Fairness in Algorithmic Decision Making (2018, background): “2017), child maltreatment hotlines (Gillingham 2015; Chouldechova et al. 2018), and credit risk assessments (Huang, Chen, and Wang 2007).”
Ensuring Fairness in Machine Learning to Advance Health Equity (2018, background): “Similar biases have been observed in predictive policing (4) and identifying which calls to a child protective services agency required an in-person investigation (5, 6).”
Understanding the Effect of Accuracy on Trust in Machine Learning Models (2019, background): “Police departments have used ML systems to predict the location of human trafficking hotspots [28], while child welfare workers have used predictive modeling to strategically target services to the children most at risk [3].”
Predictive Inequity in Object Detection (2019, background): “Examples of the latter include predictive policing, recidivism prediction, and the allocation of social welfare resources (Rubin, 2010; Maloof, 1999; Kriegler & Berk, 2010; Chouldechova et al., 2018).”