Phase 1
Completed
-
Virtual assistants (VA) like Siri, Alexa, or Google Assistant are becoming household names - especially for families with young children. Scientific inquiry studying this user population and their intention (not) to use VAs at home, however, remains scarce.
Study 1: By bridging the Technology Acceptance Model, Uses and Gratifications theory, and the first proposition of the Differential Susceptibility to Media Effects Model, this study disentangled (1) different types of families with (2) different motivations for (3) different forms of VA-usage (i.e., parent only, child only, co-use).
Study 2: By orientating along ecological system theory and following arguments of the second digital divide, this study investigated the question: Who are those families that decide to adopt a smart speaker device and how do they differ from those who do not?
-
Study 1: We conducted a cross-sectional survey including data from 305 Dutch parents with at least one child between 3-8 years and a Google Assistant-powered smart speaker in their home.
Study 2: We conducted a cross-sectional survey on a representative sample of the Dutch population from which we filtered for Dutch parents with at least one child between 3-8 years. Data included 187 valid self-reports.
-
Study 1: We found that (1) families mostly differ along parents' digital literacy skills, frequency of VA-use, trust in technology, and preferred degree of child media-mediation. (2) Hedonic motivation was key for parents to (3) co-use the VA together with their child(ren). We highlight new pathways for the methodological and theoretical study of technology use in families. Developers can best anchor VA-application among families in aspects of enjoyment while scholars and policy makers might wish to consider additional meaningful intervention criteria for the future study and guidance of family VA-use practices.
Study 2: Families with (34%) and without a smart speaker (66%) mostly differ along their trust in technology, internet literacy, and preferred media-mediation style. Findings show a clear tendency for second-digital-divide characteristics to distinguish between families with and without a smart speaker. Interestingly, children are not yet widely granted independent access to VAs at home. This study offers guidelines for industry in terms of families’ adoption of future technologies, as well as inspires future research on meaningful smart speaker interventions for families.
Phase 2
Completed
-
Theory makes clear that the media ecosystem can shape individuals’ adoption of technology by aiding in the formation of specific attitudes towards the technology (ecological system theory, framing). So far, little has been researched about how the media, and particularly so the news media, portray Virtual Assistants (VAs). For the continued study of VA-adoption, this knowledge is however crucial to better contextualize individual use-decisions. Because of that, this study inductively uncovered how VAs have been reported about in the Dutch news media over the last decade (2011 - 2022).
-
We conducted an automated content analysis using the Analysis of Topic Model Networks (i.e., ANTMN) to analyze a corpus of 2,059 Dutch news articles. ANTMN consists of a combination of topic modelling (here: BERTopic modelling), network analysis, and the detection of communities to computationally deduct frames from media texts.
-
Our results show an increase in news attention on VAs starting in 2011, with a peak in 2019. Most recently, we noticed a clear decline in volume, which might indicate a bust period where public attention about VAs drops. In addition to that, we find two central frames through which information about VAs was presented: a frame on AI and society and one on Voice assistants and their users. Those frames represent aspects of so-called societal and personal impact framing, whereby the first presents a public issue by highlighting more direct effects for individuals, and the latter addresses larger societal consequences. Over time, we saw that the Voice assistants and their users frame often times preceded the AI and society frame. This tells us that as a new technology, VAs started out as a niche theme and over time grew more into a topic of great societal importance.
Phase 3
Completed
-
Media use in families is known to impact childrenʼs further upbringing and development. With the advent of smart speakers in many family homes, the natural question therefore arises how families concretely use these devices at home and how the technology must be further designed to meet the specific needs of families. As such, this study teaches us about different smart speaker use-types across families and their profiling along individual and contextual characteristics. With this knowledge, we can draw connections between differential susceptibilities of families and their smart speaker use-patterns to aid their empowerment in this digital age and the further human-centered development of smart home technology.
-
We use a multi-method design by combining self-reports of 370 parents about their characteristics and perceived long-term smart speaker use with data donations (N = 20) of observed smart speaker use based on donated device log-histories. For the latter, we use Port - a European data donation infrastructure (https://datadonation.eu/) that enables the safe sharing of smart speaker log-data between users and researchers.
-
Data donations: Findings show that self-reported long-term smart speaker use tends to be an underestimation of what is registered by the device. This might signal how inherently habituated smart speaker use has become until now so that users no longer recall their frequent interaction with the devices.
Self-reports: We uncover six common types of use: Infrequent/Light Users, Weekly All-round Users, Family Entertainment Users, Monthly All-round Users, Parental Assistance Users, and Power All-round Users. We also find that especially individual trust dispositions of parents and their perceptions of certain narratives in the news media context shape these different ways of smart speaker use. In short: the more trust and the higher the perceptions of personal news narratives, the more intensive the use-type.