Personal Information Disclosure via Voice Assistants: The Personalization–Privacy Paradox

Abstract

The purpose of this research is to examine the factors affecting the willingness to disclose personal information based upon the privacy calculus framework and its relation to the continued usage of the Voice Assistant (VA) devices. The information disclosure is analyzed from a dual channel benefit/risk perspective through the calculus lens. Particularly, the joint effects of the two independent paths of benefits and risks that can induce the desire to disclose personal information, along with their various antecedents are proposed and empirically tested. Data are collected from 427 respondents using Amazon Mechanical Turk (MTurk) as the platform and analyzed using a Maximum Likelihood Structural Equation Modelling (ML-SEM) technique. The analysis shows that personalized services, perceived enjoyment and perceived complementarity influence the perceived benefits, which in turn positively affects the personal information disclosure. Perceived severity and perceived control are found to have a positive effect, while perceived trust has a negative effect on the perceived risks, respectively, which in turn is negatively associated with the personal information disclosure. Compared to the perceived risks, the effect of perceived benefits on the information disclosure is found to be stronger, which motivates the users to continuously use the services of the VA devices. This work extends the current literatures on privacy concerns to the users' intention of disclosing their personal information by proposing and empirically testing six hypotheses as a part of the research model. Finally, the research implications are discussed, and suggestions are provided.

A Privacy Preservation Quality of Service (QoS) Model for Data Exposure in Android Smartphone Usage

“…In terms of developing the privacy risk model, calculus knowledge should be used because it provides a precise and accurate privacy risk score [ 28 , 29 ]. Users can understand the risks they face based on their usage behaviour.…”

Revealing or concealing: privacy information disclosure in intelligent voice assistant usage- a configurational approach

“…When users enjoy the services by disclosing information, their information privacy concerns arise simultaneously. Therefore, users might be concerned that the data they shared to VA systems might be misused, be tracked or be monitored by the devices (Pal et al. , 2020), and thus impact the privacy boundary management behaviour.…”

Investigating the Personalization–Privacy Paradox in Internet of Things (IoT) Based on Dual-Factor Theory: Moderating Effects of Type of IoT Service and User Value

“…Meanwhile, although IoT is an environment wherein the occurrence of the P-P paradox phenomenon is highly likely, very few studies have assessed this phenomenon in relation to IoT. Additionally, most existing IoT studies have examined the paradox for a specific IoT service [3,32,33]. There are various types of IoT services, and the phenomenon related to the P-P paradox could differ depending on the service and usage purpose [25,30].…”

“…In IoT environments, people find it difficult to recognize activities related to autonomously collecting, combining, and sharing data. In such IoT environments, analyzing the big data collected can provide a personalized context-aware service that is tailored to individual preferences and situations, thereby improving human convenience [3]. Conversely, as the amount and type of information newly collected in IoT increase exponentially, privacy invasion issues are emerging [4].…”