Education is, for many educational theorists, inherently uncertain, open-ended and risky (Biesta 2013, Allert et al. 2017). Yet the algorithmic systems of today‘s datafied world increasingly prioritise ‘certainty and the promise of guaranteed outcomes’ (Zuboff 2019: 497). It is widely known that AI-powered predictive systems have high margins of error. However, the data visualisations of algorithmic systems across health, social work, policing, government and other social fields tend to visualise certainty, thus invisibilising the underlying approximations and uncertainties of both the algorithmic systems and the social settings in which these systems operate. This can have harmful consequences for people, in particular minoritised populations. For example, while software providers and policy makers assure the public that algorithmic systems merely provide suggestions, and that users make the final decisions, research has shown that civil servants and other practitioners find it difficult to override algorithmic recommendations, even in their area of expertise (Eubanks 2019, Allhutter et al. 2020). Critical algorithm studies have hence raised questions about how (valid) knowledge is produced and circulates through algorithmic systems, and how truth claims are made (Jarke et al. forthcoming). In this paper, we argue that data visualisations and their production of certainty play a crucial role in devaluing situated knowledges that embrace the inherent uncertainty and open-endedness of our world. After introducing this field of research, we analyse a corpus of data visualisations from English- and German-language predictive analytics platforms for education. We explore the extent to which they show uncertainty. Findings indicate that data visualisations of uncertainty in education are exceedingly rare. The paper then discusses the implications when educators make decisions based on these visualisations. It reflects on the dashboard construction of ‘at-risk’ students (Jarke & Macgilchrist 2021), the distribution of benefits and harms to students, and the constitution of possible futures. It discusses three moves to contest the encoding of certainty into spaces of educational uncertainty: First, increased algorithmic literacy, which, however, individualises responsibility for action and transformation with the user. Second, artistic data visualisations which highlight uncertainty, which, however, tend to remain within the same frame in which data are collected about individuals. Third, then, the paper draws on a feminist/critical perspective to propose data visualisations of uncertainty that move beyond individualised data to show, for instance, structural inequalities, and that are embedded in collective (sociotechnical) practices. The paper concludes by identifying methodological challenges and open questions for future research.

Allert, H., Asmussen, M., & Richter, C. (2017). Formen von Subjektivierung und Unbestimmtheit im Umgang mit datengetriebenen Lerntechnologien – eine praxistheoretische Position. Zeitschrift für Erziehungswissenschaft, 21(1), 142-158. Allhutter, D., Cech, F., Fischer, F., Grill, G., & Mager, A. (2020). Algorithmic Profiling of Job Seekers in Austria: How Austerity Politics Are Made Effective. Frontiers in Big Data, 3. Biesta, G. (2013). The Beautiful Risk of Education. London: Paradigm Publishers. Eubanks, V. (2018). Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor. New York: St. Martin's Press. Jarke, J., Prietl, B., Egbert, S., Boeva, Y., Heuer, H. (forthcoming). Algorithmic Regimes: Methods, Interactions, Politics. Amsterdam University Press Jarke, J., & Macgilchrist, F. (2021). Dashboard stories: How the narratives told by predictive analytics reconfigure roles, risk and sociality in education. Big Data & Society, 8(1), 1-15. https://doi.org/10.1177/20539517211025561 Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. London: Profile Books.