‘There are some things that I would never ask Alexa’–privacy work, contextual integrity, and smart speaker assistants

Abstract

When new technologies like smart speaker assistants (SSAs) enter private spaces, new threats to privacy emerge. Drawing on the concepts of privacy work [Nippert-Eng, C. E. (2010). Islands of privacy. University of Chicago Press.] and contextual integrity [Nissenbaum, H. F. (2010). Privacy in context: Technology, policy, and the integrity of social life. Stanford Law Books.], this study uses qualitative interviews to explore two questions about SSAs: (1) Which kinds of privacy work do users do?, and (2) What rationales underlie users’ privacy perceptions? We identify a variety of new types of privacy work, such as limiting the dissemination of one’s voice data to a single company, or interrupting conversations during accidental SSA activation. We also identify new privacy rationales including anticipated consequences of information leaks, the importance of users’ privacy skills and awareness, and the role of choice in whether and how to use SSAs. Based on our analysis of privacy rationales, we propose an expansion of the model of contextual integrity [Nissenbaum, H. F. (2010). Privacy in context: Technology, policy, and the integrity of social life. Stanford Law Books.] to improve our ability to understand how users perceive privacy with SSAs and other voluntarily adopted home information technologies. Furthermore, we find that even SSA users who say they have no privacy concerns actually work to protect their privacy. These results complicate previous theory which said that privacy concerns lead to protective behaviour because they suggest that the relationship between these two concepts may be reversed.