The use of personal voice assistants in the context of retail digitalization: An empirical study of consumers' attitude and privacy concerns

The objective of this study was to find which factors influence consumer’s attitude toward personal voice assistants used in the retail sector. Personal voice assistants are software programs that can control applications or perform tasks following voice commands of the user. In this paper, we analyzed the impact of perceived information control (PIC) and perceived privacy risk (PP) on the attitude (A) of consumers towards the technology. Also, we evaluate different indicators as proxies for the tested variables. Perceived information control (PIC) is the belief of consumers about their capability to control how their personal information will be used by personal voice assistant (adapted from Dinev et al., 2013). Perceived privacy risk, is the user’s belief that he will get negative consequences out of disclosing personal information into personal voice assistants (adapted from Dinev et al., 2013) To find an answer to our hypotheses, we built a structural equation model. We used the software LISREL and SPSS for the analysis. We concluded that perceived information control (PIC) has a positive and substantial effect on the attitude (A) of consumers to use the technology. Perceived anonymity (ANON) and perceived confidentiality (CONF) are good measures of perceived information control (PIC). Contrary to what we expected, perceived privacy risk (PP) was non-significant as a determinant of the attitude (A) of consumers to use the technology. Information sensitivity (IS), regulatory perception (PLAW), importance of information transparency (TR), and benefits of information disclosure (BEN) are good measures of perceived privacy risk (PP). We carried-out LISREL tests by groups, to identify differences between sub-groups of the respondents. The link between perceived privacy risk (PP) and the attitude (A) became significant for respondents unaware of the GDPR. We also carried-out a descriptive analysis on the barriers for not using personal voice assistants. The two most important barriers were ease-of-use of the technology and privacy concerns. Companies should make sure that personal voice assistants are user-friendly (i.e. easy to use). They also should reassure users about privacy protection. These results must be interpreted carefully due to limitations we encountered related to sample size and representativity of the sample. Our findings will help companies to be able to focus on what matters the most to their target groups, to select the appropriate personal voice assistant software.