Date of Award
5-2022
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Computer Science
Committee Chair/Advisor
Dr. Long Cheng
Committee Member
Dr. Nina Hubig
Committee Member
Dr. Feng Luo
Abstract
As the artificial intelligence (AI) technique improves, voice assistant (smart speaker) such as Amazon Alexa and Google Assistant are quickly, surely permeating into people's daily lives. With its powerful and convenient benefits and the circumstances that people started to stay at their home longer due to the pandemic, reliance on smart speakers has increased rapidly. But at the same time, concerns of security on smart speakers have increased.
In this thesis, we conducted an online user survey of smart speaker users with five different perspectives – 1) Users’ engagement with privacy policy; 2) Awareness of different policy requirements defined by smart speaker providers; 3) User’s concerns in multi-user environment; 4) Developer's perception of general policies; and 5) Social biases of speech recognition on smart speaker (Fairness). With the survey, we could reveal the users' expectations on smart speakers and their difficulties in several situations, developer's perception, developer’s difficulties on generating privacy policy and meeting conditions for each application. Interestingly, some participants showed satisfaction overall and had perception about both the appropriate privacy policy and policy requirements, but there were quite a few people who read the privacy policy and had a lot of concerns. All participants were asking for speaker recognition even though the VPA devices provides the feature already. Also, many users have experienced low accuracy of speech recognition and some of them have noticed that age, gender, race, and dialects have different effects on smart speakers understanding their commands.
Recommended Citation
Kim, Ha Young, "Understanding User Perceptions of Voice Personal Assistants" (2022). All Theses. 3742.
https://open.clemson.edu/all_theses/3742