Multi-Scale Ethics: Developing a Framework for Accounting for Technology-Driven Inequality in AI

• In Event: How Should STS Address Inequality? As a Subject, a (Dis)Value)? Theoretical and Empirical Perspectives - II

Abstract

Research in Science & Technology Studies and Innovation Studies has highlighted the role of digital technologies in driving social inequality (for instance, Aghion et al. 2015; Cozzins et al 2002; Eubanks 2017). Importantly this research points out that, rather than being unintended consequences or matters of poor regulation, many of these impacts are built into the technologies themselves – and need to be tackled as such (Jasanoff 2016). This raises important questions about how societal-level effects like inequality, can be taken account of during the ethical evaluation stage of technological development.

In this paper, I will give an empirical and theoretical perspective on whether the issue of technology driven inequality can be built into ethical evaluations of AI and data technologies through the concept of Multi-scale ethics, which aims to take account of the different levels or scales at which these technologies take effect. Taking the specific example of the Alan Turing Institute’s work to help develop an ethical framework for AI applications in the UK’s NHS, this paper will also reflect upon how STS ideas around inequality and coproduction can complement traditional philosophical approaches to ethics, how we can use these approaches to understand inequality and the challenges of bringing together philosophers, STS scholars and technologists to work on these issues.

Author

• Melanie Smallman, University College London