Can an intersectional approach to data make algorithms more socially responsible?

Caitlin Bentley1 and Chisenga Muyoya2
1 King’s College London

2 University of Sheffield

Abstract

The Covid-19 pandemic highlighted the crucial role of data in government and international efforts to manage and mitigate the spread of the disease. The UN [3] argued that “the importance of high-quality, comprehensive, inclusive and timely data for rigorous research and evidence-informed decision-making cannot be overstated.” Data are at the center of decision-making surrounding social systems, and are increasingly used in artificial intelligence and algorithmic governance applications. Yet, many countries failed to detect severe inequalities between diverse citizens due to their gender, ethnicity, disability, etc. There is now mounting evidence that Covid-19 has impacted on diverse populations in differentiated ways. Many governments likewise introduced policies to control the spread of Covid-19 that did not take into account the impact such measures would have on diverse citizens. For example, in the UK, before the pandemic, financial resilience was lower among Black African and Other Black families, which may have limited their capacity to manage the financial shocks of lockdown [2]. New data governance models that support pandemic mitigation and recovery need better to take into account such inequalities.

We argue that data-driven and algorithmic decision-making must include inclusive data and effective political participation by diverse citizens in order for algorithms to be more socially responsible. Yet, within urgent timescales and unfolding crises, time to involve citizens directly dissipates. Our contribution explores intersectional approaches to data as a means of embedding political participation into all aspects of a data value chain [1]. We will present our research to conceptualise this area of data practice, and discuss the policy implications of this approach for developing more robust health data systems and processes.

References

[1] Inclusive Data Charter. Unpacking Intersectional Approaches to Data. 2021. https://
www.data4sdgs.org/resources/unpacking-intersectional-approaches-data
.
[2] Office for National Statistics. Coronavirus and the social impacts on different ethnic groups in the UK. ons.gov.uk, 2020. https://www.ons.gov.uk/peoplepopulationandcommunity/culturalidentity/ethnicity/articles/coronavirusandthesocialimpactsondifferentethnicgroupsintheuk/2020.
[3] United Nations. UN Research Roadmap for the COVID-19 Recovery: Leveraging the Power of Science for a More Equitable, Resilient and Sustainable Future. New York, 2020.

Presentation