Hostname: page-component-745bb68f8f-cphqk Total loading time: 0 Render date: 2025-01-13T11:10:41.630Z Has data issue: false hasContentIssue false
  • English
  • Français

Using Artificial Intelligence to classify Jobseekers: The Accuracy-Equity Trade-off

Published online by Cambridge University Press:  08 May 2020

SAM DESIERE Open the ORCID record for SAM DESIERE [Opens in a new window]  and
LUDO STRUYVEN
SAM DESIERE
Affiliation:
HIVA, KU Leuven, 3000Leuven, Belgium, email: sam.desiere@kuleuven.be
LUDO STRUYVEN
Affiliation:
HIVA, KU Leuven, 3000Leuven, Belgium, email: ludo.struyven@kuleuven.be
Get access Rights & Permissions [Opens in a new window]

Abstract

Artificial intelligence (AI) is increasingly popular in the public sector to improve the cost-efficiency of service delivery. One example is AI-based profiling models in public employment services (PES), which predict a jobseeker’s probability of finding work and are used to segment jobseekers in groups. Profiling models hold the potential to improve identification of jobseekers at-risk of becoming long-term unemployed, but also induce discrimination. Using a recently developed AI-based profiling model of the Flemish PES, we assess to what extent AI-based profiling ‘discriminates’ against jobseekers of foreign origin compared to traditional rule-based profiling approaches. At a maximum level of accuracy, jobseekers of foreign origin who ultimately find a job are 2.6 times more likely to be misclassified as ‘high-risk’ jobseekers. We argue that it is critical that policymakers and caseworkers understand the inherent trade-offs of profiling models, and consider the limitations when integrating these models in daily operations. We develop a graphical tool to visualize the accuracy-equity trade-off in order to facilitate policy discussions.

Keywords

profilingstatistical discriminationpublic employment servicesartificial intelligenceVDAB
Type
Article
Information
Journal of Social Policy , Volume 50 , Issue 2 , April 2021 , pp. 367 - 385
Copyright
© The Author(s), 2020. Published by Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

AlgorithmWatch (2019). Automating society: Taking stock of a automated decision making in the EU. Berlin, AlgorithmWatch in cooperation with Bertelsmann Stiftung.Google Scholar
Angwin, J., Larson, J., Mattu, S. and Kirchner, L. (2016), Machine bias. There’s software used across the country to predict future criminals. And it’s biased against blacks. ProPublica.Google Scholar
Barnes, S.-A., Wright, S., Irving, P. and Deganis, I. (2015), Identification of latest trends and current developments in methods to profile jobseekers in European public employment services: final report.Google Scholar
Black, D.A., Galdo, J. and Smith, J.A. (2007), ‘Evaluating the worker profiling and reemployment services system using a regression discontinuity approach.’ American Economic Review, 97, 2, 104–107.Google Scholar
Black, D.A., Smith, J.A., Plesca, M. and Shannon, S. (2003), Profiling UI claimants to allocate reemployment services: Evidence and Recommendations for States. Final Report to United States Department of Labor.Google Scholar
Bouckaert, D., Reussens, M., Larnout, D., Heene, L., Schoonbrood, S., Claes, R., Klewais, E. and Humbeeck, G.V. (2017), ‘VDAB op koers voor een datagedreven aanpak met big data.’ Over Werk, 2, 64–69.Google Scholar
Brady, M. (2018), ‘Targeting single mothers? Dynamics of contracting Australian employment services and activation policies at the street level.’ Journal of Social Policy, 47(4), 827–845.Google Scholar
Busch, P.A., Henriksen, H.Z. and Sæbø, Ø. (2018), ‘Opportunities and challenges of digitized discretionary practices: a public service worker perspective.’ Government Information Quarterly, 35, 4, 547–556.Google Scholar
Calders, T. and Verwer, S. (2010), ‘Three naive Bayes approaches for discrimination-free classification.’ Data Mining and Knowledge Discovery, 21, 2, 277–292.Google Scholar
Card, D., Kluve, J. and Weber, A. (2017), ‘What works? A meta analysis of recent active labor market program evaluations.’ Journal of the European Economic Association, 16, 3, 894–931.CrossRefGoogle Scholar
Cockx, B., Lechner, M., Bollens, J. (2019), ‘Priority to unemployed immigrants? A causal machine Learning evaluation of training in Belgium.’ IZA Discussion Paper No 12875.Google Scholar
Corbett-Davies, S., Pierson, E., Feller, A., Goel, S. and Huq, A. (2017), Algorithmic decision making and the cost of fairness. Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ACM.Google Scholar
Danneels, L. and Viaene, S. (2015), ‘Simple rules strategy to transform government: An ADR approach.’ Government Information Quarterly, 32, 4, 516–525.Google Scholar
Desiere, S., Langenbucher, K. and Struyven, L. (2019), Statistical profiling in public employment services. OECD Working Paper.Google Scholar
Devlieghere, J., Bradt, L. and Roose, R. (2019), ‘Electronic information systems as means for accountability: why there is no such thing as objectivity.’ European Journal of Social Work, 1–12.Google Scholar
De Wilde, M. and Marchal, S. (2019), ‘Weighing up work willingness in social assistance: a balancing act on multiple levels.’ European Sociological Review, 35(5), 718–737.Google Scholar
Dressel, J. and Farid, H. (2018), ‘The accuracy, fairness, and limits of predicting recidivism.’ Science Advances, 4, 1.Google ScholarPubMed
Dusseldorp, E., Hofstetter, H. and Sonke, C. (2018), ‘Landelijke doorontwikkeling van de UWV Werkverkenner: eindrapportage.’ Google Scholar
Eberts, R.W., O’Leary, C.J. and Wandner, S.A. (2002), Targeting employment services, WE Upjohn Institute.Google Scholar
Eubanks, V. (2018), Automating inequality: How high-tech tools profile, police, and punish the poor, St. Martin’s Press.Google Scholar
Fletcher, D. R. (2011), ‘Welfare reform, Jobcentre Plus and the street-level bureaucracy: towards inconsistent and discriminatory welfare for severely disadvantaged groups?’ Social policy and Society, 10(4), 445–458.CrossRefGoogle Scholar
Friedler, S.A., Scheidegger, C. and Venkatasubramanian, S. (2016), ‘On the (im) possibility of fairness.’ arXiv preprint arXiv:1609.07236.Google Scholar
Goodman, B. and Flaxman, S. (2016), ‘European Union regulations on algorithmic decision-making and a “right to explanation”.’ arXiv preprint arXiv:1606.08813.Google Scholar
Hasluck, C. (2008), The use of statistical profiling for targeting employment services: The international experience. New European Approaches to Long-Term Unemployment: What role for public employment services and what market for private stakeholders.Google Scholar
Henman, P. (2004). ‘Targeted! Population segmentation, electronic surveillance and governing the unemployed in Australia.’ International Sociology, 19(2), 173–191.CrossRefGoogle Scholar
Kim, G.-H., Trimi, S. and Chung, J.-H. (2014), ‘Big-data applications in the government sector.’ Communications of the ACM, 57, 3, 78–85.Google Scholar
Kleinberg, J., Mullainathan, S. and Raghavan, M. (2016), ‘Inherent trade-offs in the fair determination of risk scores.’ arXiv preprint arXiv:1609.05807.Google Scholar
Klievink, B., Romijn, B.-J., Cunningham, S. and de Bruijn, H. (2017), ‘Big data in the public sector: Uncertainties and readiness.’ Information Systems Frontiers, 19, 2, 267–283.Google Scholar
Lechner, M. and Smith, J. (2007), ‘What is the value added by caseworkers?’ Labour Economics, 14(2), 135–151.CrossRefGoogle Scholar
Le Grand, J. (1990), ‘Equity versus efficiency: the elusive trade-off.’ Ethics, 100(3), 554–568.CrossRefGoogle Scholar
Loxha, A. and Morgandi, M. (2014), Profiling the unemployed: a review of OECD experiences and implications for emerging economies. Social Protection and labor discussion paper. World Bank Group, Washington, DC.Google Scholar
Ludwig-Mayerhofer, W., Behrend, O. and Sondermann, A. (2014). ‘Activation, public employment services and their clients: the role of social class in a continental welfare state’. Social Policy & Administration, 48, 5, 594–612.Google Scholar
OECD (1998), Early identification of jobseekers at risk of long-term unemployment: the role of profiling, OECD.Google Scholar
OECD (2005), OECD Employment Outlook, OECD.Google Scholar
Okun, A. (1975). ‘Equality and Efficiency: The Big Tradeoff.’ Washington: Brookings Institution Press.Google Scholar
Marks, M., (2019), ‘Artificial Intelligence based suicide prediction.’ Yale Journal of Health Policy, Law, and Ethics, Forthcoming.Google Scholar
Martin, K. (2018), ‘Ethical implications and accountability of algorithms.’ Journal of Business Ethics, 1–16.Google Scholar
Pope, D.G. and Sydnor, J.R. (2011), ‘Implementing anti-discrimination policies in statistical profiling models.’ American Economic Journal: Economic Policy, 3, 3, 206–231.Google Scholar
Romei, A. and Ruggieri, S. (2014), ‘A multidisciplinary survey on discrimination analysis.’ The Knowledge Engineering Review, 29, 5, 582–638.Google Scholar
Schwab, S. (1986), ‘Is statistical discrimination efficient?’ The American Economic Review, 76, 1, 228–234.Google Scholar
Shorey, S. and Howard, P. (2016). ‘Automation, big data and politics: A research review.’ International Journal of Communication 10.Google Scholar
Struyven, L. and Van Parys, L. (2014). ‘Revisiting the Pillars of the PES and Common Challenges.’ In Leroy, F. and Struyven, L. (eds.), Building Bridges. Shaping the Future of Public Employment Services Towards 2020, 49–69. Brugge: die Keure.Google Scholar
Veale, M. and Brass, I. (2019), Administration by algorithm? Public management meets public sector machine learning. Public Management Meets Public Sector Machine Learning.Google Scholar
Wijnhoven, M. and Havinga, H. (2014), ‘The Work Profiler: A digital instrument for selection and diagnosis of the unemployed.’ Local Economy, 29, 6-7, 740–749.Google Scholar
Williams, B.A., Brooks, C.F. and Shmargad, Y. (2018), ‘How algorithms discriminate based on data they lack: challenges, solutions, and policy implications.’ Journal of Information Policy, 8, 78–115.CrossRefGoogle Scholar
Wirtz, B.W., Weyerer, J.C. and Geyer, C. (2018), ‘Artificial Intelligence and the public sector—applications and challenges.’ International Journal of Public Administration, 1–20.Google Scholar
Žliobaitė, I. and Custers, B. (2016), ‘Using sensitive personal data may be necessary for avoiding discrimination in data-driven decision models.’ Artificial Intelligence and Law, 24, 2, 183–201.Google Scholar
Žliobaitė, I. (2017), ‘Measuring discrimination in algorithmic decision making.’ Data Mining and Knowledge Discovery, 31, 4, 1060–1089.Google Scholar
Supplementary material: File

Desiere and Struyven supplementary material

Desiere and Struyven supplementary material 1

Download Desiere and Struyven supplementary material(File)
File 292.4 KB
Supplementary material: Image

Desiere and Struyven supplementary material

Desiere and Struyven supplementary material 2

Download Desiere and Struyven supplementary material(Image)
Image 8.8 MB