User Perceptions and Privacy Information in the Smart Speaker Onboarding Process

Abstract

A smart speaker's onboarding process is extremely important because it is the user's first touchpoint with the device. An effective onboarding process will communicate the necessary information a user needs in order to understand the smart speaker's functionalities as well as how to interact with it. Despite this, the onboarding process has not been a properly considered channel for conveying privacy information to users. This is surprising given that recommendations for communicating privacy information and related privacy controls often include making this information salient to the user and providing instructional use for controls. In this thesis, I explore the onboarding process for smart speakers as a potentially effective medium for which to convey privacy information. I conducted an empirical assessment of smart speakers current privacy practices and their onboarding flows in order to determine where privacy information and the communication of this information may be improved. I used my findings from this analysis to develop a smart speaker prototype to test the effectiveness of the speaker's onboarding process in helping users understand the speaker's functionalities. I also designed privacy-oriented voice commands to test within the onboarding process to evaluate if this type of privacy control influences user comprehension of privacy information. The results of this thesis show that the smart speaker's onboarding process can be improved to help users understand the device's privacy practices. Furthermore, they demonstrate that privacy-oriented voice commands show potential for future research despite being ineffective in this study.