Social Robotic Surveillance and Manipulation: How Pervasive Data Collection, Personalisation, and Robotic Influence Could Harm Robot Users and Society

In this work, we examine the potential harms of social robotic surveillance and manipulation to robot users and society. We discuss these potential harms through the lenses of the sociological theories of surveillance capitalism as detailed by Zuboff, and governmentality as proposed by Foucault. To map and address these risks, we contribute: 1) An initial multilevel theoretical framework of harms on the micro-, meso- and macro-levels of society, 2) A mapping of how specific social robotic capabilities can induce harms via surveillance and manipulation, 3) The exemplification of potential harms to robot users and society through speculative Human-Robot Interaction (HRI) scenarios grounded in previous HRI literature, and 4) Recommendations on how the HRI community could begin addressing these harms. This paper aims to demonstrate that while many user-aligned and helpful use cases have been explored for social robots, the current economic paradigm of surveillance capitalism, and the aims for governmentality, can overshadow these use cases. We argue that roboticists should think critically about the prevailing economic and political systems which their robots will inhabit, consider how power is distributed in HRI, reflect on how the identified harms may be perpetuated by their research, and critically consider how these harms should be addressed.