User preferences for privacy features in digital assistants

  • PDF / 827,017 Bytes
  • 16 Pages / 595.276 x 790.866 pts Page_size
  • 0 Downloads / 196 Views

DOWNLOAD

REPORT


RESEARCH PAPER

User preferences for privacy features in digital assistants Frank Ebbers 1

&

Jan Zibuschka 2 & Christian Zimmermann 2 & Oliver Hinz 3

Received: 5 March 2020 / Accepted: 27 October 2020 # The Author(s) 2020

Abstract Digital assistants (DA) perform routine tasks for users by interacting with the Internet of Things (IoT) devices and digital services. To do so, such assistants rely heavily on personal data, e.g. to provide personalized responses. This leads to privacy concerns for users and makes privacy features an important component of digital assistants. This study examines user preferences for three attributes of the design of privacy features in digital assistants, namely (1) the amount of information on personal data that is shown to the user, (2) explainability of the DA’s decision, and (3) the degree of gamification of the user interface (UI). In addition, it estimates users’ willingness to pay (WTP) for different versions of privacy features. The results for the full sample show that users prefer to understand the rationale behind the DA’s decisions based on the personal information involved, while being given information about the potential impacts of disclosing specific data. Further, the results indicate that users prefer to interact with the DA’s privacy features in a serious game. For this product, users are willing to pay €21.39 per month. In general, a playful design of privacy features is strongly preferred, as users are willing to pay 23.8% more compared to an option without any gamified elements. A detailed analysis identifies two customer clusters “Best Agers” and “DA Advocates”, which differ mainly in their average age and willingness to pay. Further, “DA Advocates” are mainly male and more privacy sensitive, whereas “Best Agers” show a higher affinity for a playful design of privacy features. JEL classification O330 Keywords Information privacy . Digital assistant . Intelligent personal assistant . Choice-based conjoint analysis . Privacy preferences . Internet of things

Introduction Responsible Editor: Robert Harmon * Frank Ebbers [email protected] Jan Zibuschka [email protected] Christian Zimmermann [email protected] Oliver Hinz [email protected] 1

Fraunhofer Institute for Systems and Innovation Research ISI, Breslauer Str. 48, 76139 Karlsruhe, Germany

2

Robert Bosch GmbH, Robert-Bosch-Campus 1, 71272 Renningen, Germany

3

Goethe University Frankfurt, Theodor-W.-Adorno-Platz 4, 60323 Frankfurt am Main, Germany

The Internet of Things (IoT) offers a plethora of new possibilities in society and the economy. Digital assistants (DA) are one way to support users to adopt IoT applications and manage the ever increasing number of interconnected devices (Maedche et al. 2019). One possible application of a DA, for example, is that a DA automatically sets the house heating system to the user’s preferred temperature as soon as he or she is on the way home. To provide comprehensive user assistance, the DA needs to process a large amo