CEU Electronic Theses and Dissertations, 2020
| Author | Cohen, Daniel |
|---|---|
| Title | In Machines We Trust: Political Legitimacy and the Case of Automated Border Agents at Europe's Borders |
| Summary | In this essay, I question the relationship between political legitimacy and the use of Automated Decision Making (ADM) systems for public service provision. Although algorithmic governance (AGOV) mechanisms are still in their infancy, they indicate a trend of increased cognitive outsourcing to efficient and cost-effective AI Assistants. Increased automation throughout the public service changes the dynamic between the citizen and the political authority. This paper considers the ways by which political legitimacy can be measured to develop policies that build trust in algorithmic governance. With the assistance of expert interviews from AI policy practitioners, I build on a framework of political legitimacy to assess the input, output and throughput legitimacy of ADM systems. The framework is applied to the case of the Horizon 2020 pilot project iBorderCtrl, which uses Machine Learning (ML) to assist border control guards with a risk assessment of incoming travelers. I find that a non-binary approach to legitimacy, which includes the black-box of decision making, offers new avenues to study the ways by which throughput legitimacy can offer solutions to both the input of citizens’ voices into the political decision-making process as well as fair, equal and socially beneficial outcomes. The paper concludes by offering mitigation strategies at different levels of the algorithmic development cycle to address potential legitimacy issues. |
| Supervisor | Rippon, Simon; Gallego, Aina |
| Department | Public Policy MA |
| Full text | https://www.etd.ceu.edu/2020/cohen_daniel.pdf |
Visit the CEU Library.
© 2007-2021, Central European University