The Devil is in the Details: Interrogating Values Embedded in the Allegheny Family Screening Tool
Open accessThe DOI matches its Crossref record; 31 of the 31 references it cites were found in a registry or an index.
Abstract
The design decisions of developers and researchers in creating algorithmic tools — like constructing variables, performing feature selection, and binning model outputs — are sometimes cast as objective technical processes. In reality, these decisions are far from objective, and they are sometimes even made arbitrarily. In this work, we examine how algorithmic design choices can function as policy decisions through an audit of a deployed algorithmic tool, the Allegheny Family Screening Tool (AFST), used to screen calls to a child welfare agency about alleged child neglect in Allegheny County, Pennsylvania. We analyze design decisions in the AFST’s development process related to feature selection, data collection, and post-processing, highlighting three values implicitly embedded in the tool through these decisions. By aggregating risk scores at the household level, the AFST effectively treats families as “risky” by association. In choosing to use training data from the criminal legal system and behavioral health agencies, the AFST prioritizes “making decisions based on as much information as possible,” even when that information is potentially biased across race, disability, and other protected statuses. Finally, by including static features in the model that identify whether a person has ever been affected by the criminal legal system or relied on public benefits, the AFST chooses to mark families in perpetuity, compounding the impacts of systemic discrimination and foreclosing opportunities for recourse for families impacted by the tool. We explore the impacts of these decisions, individually and together, arguing that they function as policy choices that may have discriminatory effects and raise concerns about lack of democratic oversight.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
A survey on feature selection methods
Chandrashekar, G., & Sahin, F. (2014) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2013.11.024 · Cited by 5,189 · cited 1×
“There are many different ways to perform feature selection, which is often focused on ensuring that only those variables that allow a model to perform best on a performance metric decided on by the model’s developers are kept in the model [17, 53, 54].”
BackgroundAUC: A misleading measure of the performance of predictive distribution models
Lobo, J. M., Jiménez‐Valverde, A., & Real, R. (2008) · Global Ecology and Biogeography · doi:10.1111/j.1466-8238.2007.00358.x · Cited by 3,499
Model Cards for Model Reporting
Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., Spitzer, E., Raji, I. D., & Gebru, T. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287596 · Cited by 2,681 · cited 2×
“Since2016,severaladditionaliterationsofthe toolhavebeendeveloped[20,75,98];furtherdetailsaboutthe designofthetoolandchangestothetoolbetweeniterationsare describedinamodelcard[64]wedevelopedfortheAFST,included asAppendixC.”
MethodFairness and Abstraction in Sociotechnical Systems
Selbst, A. D., Boyd, D., Friedler, S. A., Venkatasubramanian, S., & Vertesi, J. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287598 · Cited by 1,631 · cited 1×
“When developers cast these kinds of design decisions as primarily technical questions [38, 92], they may disguise”
BackgroundFriedman, B. (1996) · Interactions · doi:10.1145/242485.242493 · Cited by 966 · cited 1×
“More broadly, efforts to interrogate technical systems and methods as value-laden artifacts, including in the context of use by government agencies, extend back several decades, spanning philosophy, information science, human-computer interaction, and other fields [30, 31, 61, 69].”
BackgroundDecolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence
Mohamed, S., Png, M.-T., & Isaac, W. (2020) · Philosophy & Technology · doi:10.1007/s13347-020-00405-8 · Cited by 912 · cited 1×
“[66] propose new frameworks that focus instead on connecting our understanding of algorithmic oppression to the broader social and economic contexts in which algorithms operate to evaluate whether algorithms can actually be designed to promote justice [38, 66].”
BackgroundBender, E. M., & Friedman, B. (2018) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00041 · Cited by 854 · cited 1×
“Viewing design decisions as policy choices, a related body of work focuses on the actions and functions of data scientists, engineers, researchers, and other actors involved in the design of algorithmic systems, including in shaping the datasets used to build machine learning (ML) systems [11, 67, 68, 77], defining th…”
BackgroundAlgorithmic Fairness: Choices, Assumptions, and Definitions
Mitchell, S., Potash, E., Barocas, S., D'Amour, A., & Lum, K. (2021) · Annual Review of Statistics and Its Application · doi:10.1146/annurev-statistics-042720-125902 · Cited by 687 · cited 1×
“…other actors involved in the design of algorithmic systems, including in shaping the datasets used to build machine learning (ML) systems [11, 67, 68, 77], defining the measurements used to assess ML systems [44, 45, 65], and shaping how system outputs are communicated and explained [12, 14, 58].”
BackgroundPrediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions
Mitchell, S., Potash, E., Barocas, S., D'Amour, A., & Lum, K. (2018) · arXiv · doi:10.48550/arxiv.1811.07867 · Cited by 687
Explainable machine learning in deployment
Bhatt, U., Xiang, A., Sharma, S., Weller, A., Taly, A., Jia, Y., Ghosh, J., Puri, R., Moura, J. M. F., & Eckersley, P. (2020) · Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency · doi:10.1145/3351095.3375624 · Cited by 633
Cai, C. J., Winter, S., Steiner, D., Wilcox, L., & Terry, M. (2019) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3359206 · Cited by 594 · cited 1×
“As government agencies increasingly adopt predictive tools in areas ranging from healthcare [16, 36, 102], to education [24, 76], to the criminal legal system [63], and beyond, a growing body of work has focused on understanding the selection and use of algorithmic tools as encoding policy choices, including in the cr…”
BackgroundA large-scale analysis of racial disparities in police stops across the United States
Pierson, E., Simoiu, C., Overgoor, J., Corbett-Davies, S., Jenson, D., Shoemaker, A., Ramachandran, V., Barghouty, P., Phillips, C., Shroff, R., et al. (2020) · Nature Human Behaviour · doi:10.1038/s41562-020-0858-1 · Cited by 552 · cited 1×
“can increase an individual’s and household’s AFST score, raising serious concerns that including these racially disparate features — which reflect the racially biased policing and criminal legal systems [4, 9, 60, 80, 83] — could exacerbate and reify existing racial biases.”
BackgroundJacobs, A. Z., & Wallach, H. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445901 · Cited by 362 · cited 2×
“surements used to assess ML systems [44, 45, 65], and shaping how system outputs are communicated and explained [12, 14, 58].”
BackgroundA Survey of Algorithmic Recourse: Contrastive Explanations and Consequential Recommendations
Karimi, A.-H., Barthe, G., Schölkopf, B., & Valera, I. (2023) · ACM Computing Surveys · doi:10.1145/3527848 · Cited by 301
The flaws of policies requiring human oversight of government algorithms
Green, B. (2022) · Computer Law & Security Review · doi:10.1016/j.clsr.2022.105681 · Cited by 263
Government by Algorithm: Artificial Intelligence in Federal Administrative Agencies
Engstrom, D. F., Ho, D. E., Sharkey, C. M., & Cuéllar, M.-F. (2020) · SSRN Electronic Journal · doi:10.2139/ssrn.3551505 · Cited by 262
The Flaws of Policies Requiring Human Oversight of Government Algorithms
Green, B. (2021) · doi:10.2139/ssrn.3921216 · Cited by 259
Yang, Q., Steinfeld, A., & Zimmerman, J. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300468 · Cited by 247
A Case for Humans-in-the-Loop: Decisions in the Presence of Erroneous Algorithmic Scores
De-Arteaga, M., Fogliato, R., & Chouldechova, A. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376638 · Cited by 221 · cited 2×
“…consider aspects of design and deployment beyond those considered here, including evaluations of the version of the AFST currently deployed, more detailed analyses examining outcomes during tool deployment, and other evaluations covering questions discussed in [18, 19, 26, 50] and other works.”
MethodBrown, A., Chouldechova, A., Putnam-Hornstein, E., Tobin, A., & Vaithianathan, R. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300271 · Cited by 207
Kawakami, A., Sivaraman, V., Cheng, H.-F., Stapleton, L., Cheng, Y., Qing, D., Perer, A., Wu, Z. S., Zhu, H., & Holstein, K. (2022) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3491102.3517439 · Cited by 143 · cited 2×
“Much of the call screeners’ work of looking for the reasoning behind racially and socioeconomically disparate risk score outputs from the AFST model may be guesswork [49], as call screeners are not shown the precise values of features that are used to generate scores.”
BackgroundA Human-Centered Review of Algorithms used within the U.S. Child Welfare System
Saxena, D., Badillo-Urquiola, K., Wisniewski, P. J., & Guha, S. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376229 · Cited by 116 · cited 2×
“These tools can vary in their application context, training data, and outcomes; for a survey and analysis of the different types of predictive analytics tools used in jurisdictions around the country, see Samant et al. [90] and Saxena et al. [91].”
BackgroundThe philosophical basis of algorithmic recourse
Venkatasubramanian, S., & Alfano, M. (2020) · Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency · doi:10.1145/3351095.3372876 · Cited by 114 · cited 1×
“…for affected parties to receive explanations about how an algorithm generated a particular output [48], while other definitions suggest that recourse is the ability to not only know how an algorithm reaches its decisions, but also to change decisions stemming from an algorithmic system [100].”
BackgroundModel Multiplicity: Opportunities, Concerns, and Solutions
Black, E., Raghavan, M., & Barocas, S. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533149 · Cited by 85 · cited 3×
“Even taking as given a prioritization of some measure of accuracy in the model development process, model multiplicity could allow tool developers and designers to prioritize considerations beyond accuracy [13] — including interpretability, opportunity for recourse, and fairness.”
BackgroundMethodA Framework for Understanding Sources of Harm throughout the Machine Learning Life Cycle
Suresh, H., & Guttag, J. (2021) · Equity and Access in Algorithms, Mechanisms, and Optimization · doi:10.1145/3465416.3483305 · Cited by 82 · cited 2×
“In this section, we analyze this series of post-processing decisions [95] related to the aggregation and communication of the AFST’s outputs.”
BackgroundMethodA Framework for Understanding Sources of Harm throughout the Machine Learning Life Cycle
Suresh, H., & Guttag, J. V. (2019) · arXiv · doi:10.48550/arxiv.1901.10002 · Cited by 82
Procurement As Policy: Administrative Process for Machine Learning
Mulligan, D. K., & Bamberger, K. A. (2019) · SSRN Electronic Journal · doi:10.2139/ssrn.3464203 · Cited by 75 · cited 3×
“Mulligan and Bamberger [70] argue that, under the current paradigm of government agencies procuring and using algorithmic tools developed by third-party entities, government decision-makers abdicate im-”
BackgroundInfluentialRobertson, S., Nguyen, T., & Salehi, N. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445748 · Cited by 67
Constructing Recidivism Risk for Sentencing
Eaglin, J. (2016) · SSRN Electronic Journal · doi:10.2139/ssrn.2821136 · Cited by 45
Heterogeneity in Algorithm-Assisted Decision-Making: A Case Study in Child Abuse Hotline Screening
Cheng, L., & Chouldechova, A. (2022) · arXiv · doi:10.48550/arxiv.2204.05478 · Cited by 20
Cited by
Why ‘open’ AI systems are actually closed, and why this matters
Widder, D. G., Whittaker, M., & West, S. M. (2024) · Nature · doi:10.1038/s41586-024-08141-1 · Cited by 86
Kawakami, A., Coston, A., Zhu, H., Heidari, H., & Holstein, K. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642849 · Cited by 54
“In the past decade, such AI tools have often failed to serve the needs of the communities that agencies are expected to serve [6, 17, 20, 24, 25].”
BackgroundInfluentialKawakami, A., Coston, A., Heidari, H., Holstein, K., & Zhu, H. (2024) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3686989 · Cited by 24
Disclosure or Marketing? Analyzing the Efficacy of Vendor Self-reports for Vetting Public-sector AI
Kuehnert, B., Johnson, N., Dotan, R., & Heidari, H. (2026) · Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3805689.3812220
Metadata
- Authors
- Gerchick, M., Jegede, T., Shah, T., Gutierrez, A., Beiers, S., Shemtov, N., Xu, K., Samant, A., & Horowitz, A.
- Venue
- 2023 ACM Conference on Fairness Accountability and Transparency
- Publisher
- ACM
- Published
- 2023-06-12
- Pages
- 1292-1310
- OpenAlex
- W4380320146
- Cited by
- 34
- References
- 31
- FWCI
- 8.77
- Type
- conference paper
- Language
- en
- Open access
- bronze
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.