Fairer machine learning in the real world: Mitigating discrimination without collecting sensitive data
Open accessPublished version of a preprint
The DOI matches its Crossref record; 58 of the 58 references it cites were found in a registry or an index.
Abstract
Decisions based on algorithmic, machine learning models can be unfair, reproducing biases in historical data used to train them. While computational techniques are emerging to address aspects of these concerns through communities such as discrimination-aware data mining (DADM) and fairness, accountability and transparency machine learning (FATML), their practical implementation faces real-world challenges. For legal, institutional or commercial reasons, organisations might not hold the data on sensitive attributes such as gender, ethnicity, sexuality or disability needed to diagnose and mitigate emergent indirect discrimination-by-proxy, such as redlining. Such organisations might also lack the knowledge and capacity to identify and manage fairness issues that are emergent properties of complex sociotechnical systems. This paper presents and discusses three potential approaches to deal with such knowledge and information deficits in the context of fairer machine learning. Trusted third parties could selectively store data necessary for performing discrimination discovery and incorporating fairness constraints into model-building in a privacy-preserving manner. Collaborative online platforms would allow diverse organisations to record, share and access contextual and experiential knowledge to promote fairness in machine learning systems. Finally, unsupervised learning and pedagogically interpretable algorithms might allow fairness hypotheses to be built for further selective testing and exploration. Real-world fairness challenges in machine learning are not abstract, constrained optimisation problems, but are institutionally and contextually grounded. Computational fairness tools are useful, but must be researched and developed in and with the messy contexts that will shape their deployment, rather than just for imagined situations. Not doing so risks real, near-term algorithmic harm.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“The Article 22 prohibition is limited to decisions which produce legal or similarly significant effects concerning the data subject 41 .”
From Administrative law and the machines of government: judicial review of automated public-sector decision-making (2019) · cites this paper as [41]
In the citing paper
“Many of the methods to tackle discrimination in AI systems implicitly assume that organisations hold these sensitive data”
From Discrimination, artificial intelligence, and algorithmic decision-making (2025) · cites this paper as Veale and Binns 2017
In the citing paper
“Not considering correlated variables has been shown to increase the risk of discrimination (e.g., redlining 4 ) [54,88,89,89,196,202,231,247,286,308].”
From Fairness in machine learning: A survey (2024) · cites this paper as [286]
In the citing paper
“Computing by hand can introduce new biases without training [286], as fairness is typically not monotonic, thus assigning a threshold value may be quite arbitrary [270].”
From Fairness in machine learning: A survey (2024) · cites this paper as [286]
In the citing paper
“However, beyond this, critics of current research [50,193,266,286,287,297,297,306] argue that eforts will fail unless contextual, sociocultural, and social policy challenges are better understood.”
From Fairness in machine learning: A survey (2024) · cites this paper as [286]
4 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
DiMaggio, P. J., & Powell, W. W. (1983) · American Sociological Review · doi:10.2307/2095101 · Cited by 38,435
"Why Should I Trust You?": Explaining the Predictions of Any Classifier
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939778 · Cited by 17,555
Sorting Things Out: Classification and Its Consequences
Bowker, G. C., & Star, S. L. (1999) · doi:10.7551/mitpress/6352.001.0001 · Cited by 6,188
boyd, D., & Crawford, K. (2012) · Information, Communication & Society · doi:10.1080/1369118x.2012.678878 · Cited by 5,479
Pandora's Hope: Essays on the Reality of Science Studies
Casper, M. J., & Latour, B. (2000) · Contemporary Sociology · doi:10.2307/2655272 · Cited by 5,049
Dwork, C., Hardt, M., Pitassi, T., Reingold, O., & Zemel, R. (2012) · Proceedings of the 3rd Innovations in Theoretical Computer Science Conference · doi:10.1145/2090236.2090255 · Cited by 3,501
Wikidata: a free collaborative knowledgebase
Vrandečić, D., & Krötzsch, M. (2014) · Communications of the ACM · doi:10.1145/2629489 · Cited by 3,342
How the machine “thinks”: Understanding opacity in machine learning algorithms
Burrell, J. (2016) · Big Data & Society · doi:10.1177/2053951715622512 · Cited by 3,196 · cited 2×
“These are nested within a range of linked concerns, including algorithmic transparency and accountability (Burrell, 2016; Keats Citron and Pasquale, 2014; Kroll et al., 2016; Nissenbaum, 1996), in-the-wild reliability (Žliobaite_ et al.”
BackgroundBarocas, S., & Selbst, A. D. (2014) · doi:10.2139/ssrn.2477899 · Cited by 3,048 · cited 1×
“Some harms surround fairness, as it proves to be diffi-cult to make systems that do not exhibit bias, indirectly or in subsets of data (Hajian, 2013; Kamiran et al., 2012; Barocas and Selbst, 2016).”
BackgroundSemantics derived automatically from language corpora contain human-like biases
Caliskan, A., Bryson, J. J., & Narayanan, A. (2017) · Science · doi:10.1126/science.aal4230 · Cited by 2,887
Equality of Opportunity in Supervised Learning
Hardt, M., Price, E., & Srebro, N. (2016) · arXiv · doi:10.48550/arxiv.1610.02413 · Cited by 2,536 · cited 4×
“…‘equality of opportunity’, which considers whether each group is equally likely to be predicted a desirable outcome given the actual base rates for that group (Hardt et al., 2016); and ‘disparate mistreatment’, a corollary which considers differences in false positive rates between groups (Zafar et…”
BackgroundInfluentialFair prediction with disparate impact: a study of bias in recidivism prediction instruments
Chouldechova, A. (2017) · Big Data · doi:10.1089/big.2016.0047 · Cited by 2,159 · cited 1×
“But unfortunately, recent work has formally proven that it is impossible for a model to satisfy several of these constraints at the same time, except in exceptional cases which are unlikely to hold in the real world (Berk et al., 2017; Chouldechova, 2017; Kleinberg et al., 2016).”
BackgroundCertifying and Removing Disparate Impact
Feldman, M., Friedler, S. A., Moeller, J., Scheidegger, C., & Venkatasubramanian, S. (2015) · Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2783258.2783311 · Cited by 1,827
Dataset Shift in Machine Learning
(2008) · doi:10.7551/mitpress/9780262170055.001.0001 · Cited by 1,553 · cited 2×
“Concept drift or dataset shift refers to either real or virtual (differently sampled) changes in the conditional distributions of model inputs and outputs (Quiñonero-Candela et al., 2009) – for example, how changes in law might qualitatively affect prison population or the strategies of fraudsters.”
BackgroundExplaining nonlinear classification decisions with deep taylor decomposition
Montavon, G., Lapuschkin, S., Binder, A., Samek, W., & Müller, K.-R. (2017) · Pattern Recognition · doi:10.1016/j.patcog.2016.11.008 · Cited by 1,357 · cited 1×
“This method is not exclusive to pedagogical methods, and some recent work has shown how decompositional meth-ods, which use components of model structure rather than just treating it like a black box, also display strong promise in this space (Montavon et al., 2017).”
MethodSurvey and critique of techniques for extracting rules from trained artificial neural networks
Andrews, R., Diederich, J., & Tickle, A. B. (1995) · Knowledge-Based Systems · doi:10.1016/0950-7051(96)81920-4 · Cited by 1,225
Huang, L., Joseph, A. D., Nelson, B., Rubinstein, B. I. P., & Tygar, J. D. (2011) · Proceedings of the 4th ACM workshop on Security and artificial intelligence · doi:10.1145/2046684.2046692 · Cited by 1,194 · cited 2×
“…Citron and Pasquale, 2014; Kroll et al., 2016; Nissenbaum, 1996), in-the-wild reliability (Z ˇ liobaite _ et al., 2016); security against adversaries (Huang et al., 2011; McDaniel et al., 2016); entrenchment of inequality (boyd and Crawford, 2012; Harcourt, 2006); risks to privacy and due process…”
BackgroundKusner, M. J., Loftus, J. R., Russell, C., & Silva, R. (2017) · arXiv · doi:10.48550/arxiv.1703.06856 · Cited by 995 · cited 1×
“…relation to true/false positives/negatives, but on counterfactual scenarios wherein members of the protected groups are instead members of the non-protected group (i.e. a woman classified by the system should get the same classification she would have done had she been a man) (Kusner et al., 2017).”
BackgroundCode/Space: Software and Everyday Life
Kitchin, R., & Dodge, M. (2011) · doi:10.7551/mitpress/9780262042482.001.0001 · Cited by 964
Inherent Trade-Offs in the Fair Determination of Risk Scores
Kleinberg, J., Mullainathan, S., & Raghavan, M. (2016) · arXiv · doi:10.48550/arxiv.1609.05807 · Cited by 941
The Logic of Collective Action: Public Goods and the Theory of Groups
Midgley, L., & Olson, M. (1969) · The Western Political Quarterly · doi:10.2307/446187 · Cited by 920
Evaluating Learning Algorithms: A Classification Perspective
Japkowicz, N., & Shah, M. (2011) · doi:10.1017/cbo9780511921803 · Cited by 767
Model-Agnostic Interpretability of Machine Learning
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016) · arXiv · doi:10.48550/arxiv.1606.05386 · Cited by 699 · cited 2×
“Pedagogical interpretation has recently been described as ‘model-agnostic’ interpretation (Ribeiro et al., 2016a).”
BackgroundDoes automation bias decision-making?
SKITKA, L. J., MOSIER, K. L., & BURDICK, M. (1999) · International Journal of Human-Computer Studies · doi:10.1006/ijhc.1999.0252 · Cited by 694
Predictive Policing: The Role of Crime Forecasting in Law Enforcement Operations
Perry, W., McInnis, B., Price, C., Smith, S., & Hollywood, J. (2013) · doi:10.7249/rr233 · Cited by 693
Discrimination-aware data mining
Pedreshi, D., Ruggieri, S., & Turini, F. (2008) · Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/1401890.1401959 · Cited by 676 · cited 1×
“Even though this research field has some history (Andrews et al., 1995; Custers et al., 2013; Hajian, 2013; Pedreshi et al., 2008; Vedder, 1999), usable software libraries remain largely unavailable, and little training exists.”
BackgroundEvidence-Based Policy: A Practical Guide to Doing It Better
Cartwright, N., & Hardie, J. (2012) · doi:10.1093/acprof:osobl/9780199841608.001.0001 · Cited by 672
Elements of Statistical Disclosure Control
Willenborg, L., & de Waal, T. (2001) · Lecture Notes in Statistics · doi:10.1007/978-1-4613-0121-9 · Cited by 649
Zafar, M. B., Valera, I., Gomez Rodriguez, M., & Gummadi, K. P. (2017) · Proceedings of the 26th International Conference on World Wide Web · doi:10.1145/3038912.3052660 · Cited by 632 · cited 2×
“…opportunity’, which considers whether each group is equally likely to be predicted a desirable outcome given the actual base rates for that group (Hardt et al., 2016); and ‘disparate mistreatment’, a corollary which considers differences in false positive rates between groups (Zafar et al., 2016).”
BackgroundWe Need Both Exploratory and Confirmatory
Tukey, J. W. (1980) · The American Statistician · doi:10.1080/00031305.1980.10482706 · Cited by 571
The threat of algocracy: reality, resistance and accommodation
Danaher, J. (2016) · Philosophy & Technology · doi:10.1007/s13347-015-0211-1 · Cited by 519 · cited 2×
“…between organisations deploying such systems and those affected by them, co-operation may not be feasible or desirable; affected groups may instead be drawn (understandably) to more adversarial forms of resistance and political action (Brunton and Nissenbaum, 2015; Danaher, 2016; Lyon, 2007).”
BackgroundPrinciples and procedures of exploratory data analysis.
Behrens, J. T. (1997) · Psychological Methods · doi:10.1037/1082-989x.2.2.131 · Cited by 465 · cited 1×
“Exploratory data analysis is a core part of data analysis, but teaching, research and practice into it has been historically marginalised (Behrens, 1997; Tukey, 1980).”
BackgroundTickle, A. B., Andrews, R., Golea, M., & Diederich, J. (1998) · IEEE Transactions on Neural Networks · doi:10.1109/72.728352 · Cited by 438
Statistical Disclosure Control
Hundepool, A., Domingo‐Ferrer, J., Franconi, L., Giessing, S., Nordholt, E. S., Spicer, K., & de Wolf, P. (2012) · Wiley eBooks · doi:10.1002/9781118348239 · Cited by 385
Smart Technologies and the End(s) of Law: Novel Entanglements of Law and Technology
Hildebrandt, M. (2015) · doi:10.4337/9781849808774 · Cited by 353
Accountability in a computerized society
Nissenbaum, H. (1996) · Science and Engineering Ethics · doi:10.1007/bf02639315 · Cited by 346
A Methodology for Direct and Indirect Discrimination Prevention in Data Mining
Hajian, S., & Domingo-Ferrer, J. (2013) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2012.72 · Cited by 312
(2008) · Choice Reviews Online · doi:10.5860/choice.46-2349 · Cited by 267
Obfuscation: A User's Guide for Privacy and Protest
Brunton, F., & Nissenbaum, H. (2015) · The MIT Press eBooks · doi:10.7551/mitpress/9780262029735.001.0001 · Cited by 256
Profiling the European Citizen
(2008) · Springer eBooks · doi:10.1007/978-1-4020-6914-7 · Cited by 215
Against Prediction: Profiling, Policing, and Punishing in an Actuarial Age
Harcourt, B. E. (2006) · doi:10.7208/chicago/9780226315997.001.0001 · Cited by 193
(2012) · Law, Governance and Technology Series · doi:10.1007/978-94-007-2543-0 · Cited by 183
A Review of Financial Accounting Fraud Detection based on Data Mining Techniques
Sharma, A., & Kumar Panigrahi, P. (2012) · International Journal of Computer Applications · doi:10.5120/4787-7016 · Cited by 179
Mantelero, A. (2016) · Computer Law & Security Review · doi:10.1016/j.clsr.2016.01.014 · Cited by 147
Data protection impact assessments: a meta-regulatory approach
Binns, R. (2017) · International Data Privacy Law · doi:10.1093/idpl/ipw027 · Cited by 130
Self-citationŽliobaitė, I., & Custers, B. (2016) · Artificial Intelligence and Law · doi:10.1007/s10506-016-9182-5 · Cited by 125
Handling Conditional Discrimination
liobaite, I., Kamiran, F., & Calders, T. (2011) · 2011 IEEE 11th International Conference on Data Mining · doi:10.1109/icdm.2011.72 · Cited by 125
Children in the Public Benefit System at Risk of Maltreatment
Vaithianathan, R., Maloney, T., Putnam-Hornstein, E., & Jiang, N. (2013) · American Journal of Preventive Medicine · doi:10.1016/j.amepre.2013.04.022 · Cited by 122 · cited 1×
“…anticipating the geospatial distribution of crime (Azavea, 2015; Perry et al., 2013; Wetenschappelijke Raad voor het Regeringsbeleid (WRR), 2016), the need for child protection (Vaithianathan et al., 2013) and the detection of tax fraud (Khwaja et al., 2011; Sharma and Kumar Panigrahi, 2012).”
BackgroundCollaborative ontology engineering: a survey
Simperl, E. P. B., & Luczak–Roesch, M. (2013) · The Knowledge Engineering Review · doi:10.1017/s0269888913000192 · Cited by 110
KDD: The challenge to individualism
Vedder, A. (1999) · Ethics and Information Technology · doi:10.1023/a:1010016102284 · Cited by 102
Discrimination and Privacy in the Information Society: Data Mining and Profiling in Large Databases
(2013) · Studies in Applied Philosophy, Epistemology and Rational Ethics · doi:10.1007/978-3-642-30487-3 · Cited by 100
Law, Human Agency and Autonomic Computing
(2011) · Routledge eBooks · doi:10.4324/9780203828342 · Cited by 81
Hajian, S., Domingo‐Ferrer, J., & Farràs, O. (2014) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-014-0346-1 · Cited by 73
A framework for enhancing the social good in computing education: a values approach
Goldweber, M., Barr, J., Clear, T., Davoli, R., Mann, S., Patitsas, E., & Portnoff, S. (2013) · ACM Inroads · doi:10.1145/2432596.2432616 · Cited by 65
Enhancing the social issues components in our computing curriculum: computing for the social good
Goldweber, M., Davoli, R., Little, J. C., Riedesel, C., Walker, H., Cross, G., & Von Konsky, B. R. (2011) · ACM Inroads · doi:10.1145/1929887.1929907 · Cited by 57
Finerman, A. (1969) · ACM Computing Surveys · doi:10.1145/356540.356541 · Cited by 44
Khwaja, M. S., Awasthi, R., & Loeprick, J. (2011) · The World Bank Open Knowledge Repository (World Bank) · doi:10.1596/978-0-8213-8754-2 · Cited by 40
Synchronous digital circuits as functional programs
Gammie, P. (2013) · ACM Computing Surveys · doi:10.1145/2543581.2543588 · Cited by 28
Research works printed but not linked: 10
- Angwin J, 2016, ProPublica, 23
- Barocas S, 2016, California Law Review, 104, 671
- Hacking I, 1995, Causal Cognition: A Multidisciplinary Debate, 351
- Kamiran F, 2012, Discrimination and Privacy in the Information Society, 223
- Keats Citron D, 2014, Washington Law Review, 89, 1
- Kroll JA, 2016, University of Pennsylvania Law Review, 165, 633
- Lyon D (2007) Resisting Surveillance. The Surveillance Studies Reader, London: McGraw-Hill Education, pp. 368–377.
- Pasquale F, 2010, Northwestern University Law Review, 104, 105
- Sandvig C, 2015, Media-N, 11, 35
- van der Sloot B, 2012, Discrimination and Privacy in the Information Society, 273
Cited by
The global landscape of AI ethics guidelines
Jobin, A., Ienca, M., & Vayena, E. (2019) · Nature Machine Intelligence · doi:10.1038/s42256-019-0088-2 · Cited by 6,220
The Ethics of AI Ethics -- An Evaluation of Guidelines
Hagendorff, T. (2019) · arXiv · doi:10.48550/arxiv.1903.03425 · Cited by 1,907
“Many of those endeavors are unified under the FAT ML or XAI community (Veale and Binns 2017; Selbst et al. 2018).”
The Ethics of AI Ethics: An Evaluation of Guidelines
Hagendorff, T. (2020) · Minds and Machines · doi:10.1007/s11023-020-09517-8 · Cited by 1,911
“Many of those endeavors are unified under the FAT ML or XAI community (Veale and Binns 2017; Selbst et al. 2018).”
Improving Fairness in Machine Learning Systems: What Do Industry Practitioners Need?
Holstein, K., Wortman Vaughan, J., Daumé, H., Dudik, M., & Wallach, H. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300830 · Cited by 777
“Recent work in the fair ML literature has proposed encryption mechanisms that ensure any collected individual-level demographics can only be used for auditing [60, 93].”
BackgroundInfluentialAlgorithmic fairness in artificial intelligence for medicine and healthcare
Chen, R. J., Wang, J. J., Williamson, D. F. K., Chen, T. Y., Lipkova, J., Lu, M. Y., Sahai, S., & Mahmood, F. (2023) · Nature Biomedical Engineering · doi:10.1038/s41551-023-01056-8 · Cited by 750
Metadata
- Authors
- Veale, M., & Binns, R.
- Affiliations
- University College London (GB)
- University of Oxford (GB)
- Corresponding author
- Veale, Michael
- Venue
- Big Data & Society
- Publisher
- SAGE Publications
- Published
- 2017-11-20
- Volume
- 4
- Issue
- 2
- Pages
- 205395171774353
- ISSN
- 2053-9517 (ISSN-L)
- OpenAlex
- W2765146466
- Semantic Scholar
- 61d37ec780aba86806af77b36d513624ec1b66b2
- Cited by
- 512
- References
- 48
- Citations by year
- 2026: 84 · 2025: 97 · 2024: 59 · 2023: 74 · 2022: 78 · 2021: 50 · 2020: 35 · 2019: 20 · 2018: 14 · 2017: 1
- FWCI
- 43.94
- Type
- journal article
- Language
- en
- Open access
- gold
- dx.doi.org
- Licence
- cc-by
- journals.sagepub.com/page/policies/text-and-data-mining-license (tdm)
- Locations
- Big Data & Society (publishedVersion, cc-by)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- UCL Discovery (University College London) (submittedVersion)
- SocArXiv (OSF Preprints) (submittedVersion, cc-by)
- OSF Preprints (OSF Preprints) (submittedVersion, cc-by)
- Funders
- Engineering and Physical Sciences Research Council (EP/J017728/2)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Keywords
- algorithmic fairness, indirect discrimination, sensitive attributes, trusted third party, collaborative platforms, unsupervised learning, interpretable algorithms, algorithmic harms, redlining, socio-technical systems
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.