Lavado Nalvaiz, Natalia
Lucia Palacios, Laura (dir.) ; Pérez López, Raúl (dir.)
Universidad de Zaragoza,
2024
Abstract: Artificial Intelligence (AI) has become a significant topic in professional and academic circles in recent years (Mehta et al., 2022; Vaid et al., 2023). Companies have started investing in AI-based technology to enhance the consumer experience, such as smart speakers or voice assistants, which have gained popularity (Klaus and Zaichowsky, 2021). The use of smart speakers has grown considerably in recent years, and the global market for these devices is expected to grow from USD 11 billion in 2022 to USD 100 billion in 2032. However, while the use of smart speakers has grown considerably, so too have users' concerns about privacy. In this sense, the Marketing Science Institute established as a research priority for the period 2022-2024 the concern for consumer privacy, as it may constitute a brake on the progress of artificial intelligence and its benefits for companies.
Smart home speakers can perform a wide variety of tasks, such as checking scheduled calendar events, playing music, ordering food, or controlling other smart home devices following a verbal or written command (Chatterjee and Karahanna, 2019; Gao and Liu, 2022). At the same time, they are used to respond to users' requests in a personalised way and to collect information about their behaviour, consumption habits, tastes and preferences, as well as the environment in which they operate (Mogaji et al., 2020; Kumar et al., 2019). For this personalised service to work properly, it is necessary to obtain personal information from users. Also, the devices are equipped with microphones that are constantly on the alert, waiting for the word or command to activate them, so they are constantly listening to the conversations around them. This raises very sensitive issues in relation to perceived surveillance and intrusiveness.
The aim of this Thesis is to study how companies can build trust and reduce the negative consequences of the perception of privacy loss due to information collection and thus emphasise the increased and continued use of smart speakers. To this end, and with the aim of contributing to the existing literature, the thesis proposes three mechanisms that can increase users' trust and affect their perceptions of privacy loss: 1) encouraging personalised services, 2) investing in greater humanisation of interactions with users, and 3) increasing transparency and perceived user control.
The first empirical study aims to analyse how the personalisation offered by smart speakers influences users' attitudes towards information collection, distinguishing between overt and covert collection. In addition, this study focuses on examining the effect of users' attitudes towards overt and covert information collection on the intention to continue using the device. The mediating role of trust in the relationship between consumer attitudes and intention to continue using the device is also examined. This study is framed by the Privacy Calculus Theory and the Personalisation-Privacy Paradox (Culnan and Armstrong, 1999; Dinev and Hart, 2006; Xu et al., 2011). It is explored whether the value of receiving personalised information can determine attitudes towards two different types of information collection by the service provider and whether it can also create a trust relationship between users and providers. The results show that the value of personalisation of messages improves users' attitudes towards both types of information collection (overt and covert). This study demonstrates the important role of trust in information management as a positive relationship is found between the value of personalisation and trust, as well as between trust and the intention to continue to use it.
The second empirical study examines whether the humanisation of smart speakers can improve users' attitudes towards covert information collection and build trust. Furthermore, it studies whether such humanisation can influence the reduction of the perception of surveillance exerted by these devices, focusing on the advantages they can exert in terms of social presence. To address these objectives, a theoretical framework is developed based on the literature on humanisation and specifically the three theories mentioned above (Realism Maximisation Theory, Uncanny Valley Theory and Parasocial Relationship Theory). The results reveal that perceived humanisation helps to increase feelings of social presence and also improves users' attitudes towards covert information gathering. It is also shown that humanisation negatively influences trust when humanisation is very low. However, there comes a point of humanisation beyond which its influence is positive. Furthermore, it is shown that humanisation has a negative impact on the perception of vigilance. Finally, this research concludes that if the user trusts the company offering the service, his or her attitude towards the collection of information will be positive, even though users may think that the information has been collected covertly.
The third empirical study analyses how notice and choice options, as tools of corporate privacy policy, can enhance trust in smart home speaker providers. In addition, it examines the role of two types of attitudes (the degree of sensitivity to personal information and the importance of information transparency) on the effect of notice and choice on the effectiveness of privacy policies and on trust. This study is based on the Privacy-Trust-Behavioural Intenttion model proposed by Liu (2005). The results confirm the Privacy-Trust-Behavioural Intent model, which suggests that notice has a direct and indirect effect on trust, mediated by privacy policy effectiveness. On the other hand, choice options have only a total indirect effect on trust. This research also confirms that the privacy policy of smart home speakers can be valued differently depending on users' characteristics or attitudes. In particular, it demonstrates the positive moderating effect of sensitivity to personal information and information transparency on the impact of notice and choice options on privacy policy effectiveness and trust.
This dissertation has several theoretical contribution. It is examined how to eliminate or mitigate two main issues of concern for users: covert information collection and perceived surveillance. It provides new evidence on how to improve user attitudes towards covert information collection and contributes to specific theories related to privacy research. We also contribute to specific theories related to privacy research. In particular, our results contribute to the Personalisation-Privacy Paradox and the Privacy-Trust-Behavioural Intention model. Finally, we contribute to existing literatura on humanisation. This thesis shows that devices with human-like characteristics can improve attitudes towards covert information collection by increasing trust in the provider, creating a greater sense of social presence, and reducing the perception of surveillance.
From a managerial perspective, companies should offer users a personalised experience, as the value of receiving personalised information has been shown to generate greater continuity of use and, more importantly, improve user attitudes. In addition, providers of these devices and services must develop robust and secure transparency, privacy and security policies that assure users that the company is making appropriate use of their personal data, regardless of how it is collected. When it comes to smart home speaker design, attention should be paid to providing them with human-like features, such as more natural language or more human-like interaction.
Abstract (other lang.):
Smart home speakers can perform a wide variety of tasks, such as checking scheduled calendar events, playing music, ordering food, or controlling other smart home devices following a verbal or written command (Chatterjee and Karahanna, 2019; Gao and Liu, 2022). At the same time, they are used to respond to users' requests in a personalised way and to collect information about their behaviour, consumption habits, tastes and preferences, as well as the environment in which they operate (Mogaji et al., 2020; Kumar et al., 2019). For this personalised service to work properly, it is necessary to obtain personal information from users. Also, the devices are equipped with microphones that are constantly on the alert, waiting for the word or command to activate them, so they are constantly listening to the conversations around them. This raises very sensitive issues in relation to perceived surveillance and intrusiveness.
The aim of this Thesis is to study how companies can build trust and reduce the negative consequences of the perception of privacy loss due to information collection and thus emphasise the increased and continued use of smart speakers. To this end, and with the aim of contributing to the existing literature, the thesis proposes three mechanisms that can increase users' trust and affect their perceptions of privacy loss: 1) encouraging personalised services, 2) investing in greater humanisation of interactions with users, and 3) increasing transparency and perceived user control.
The first empirical study aims to analyse how the personalisation offered by smart speakers influences users' attitudes towards information collection, distinguishing between overt and covert collection. In addition, this study focuses on examining the effect of users' attitudes towards overt and covert information collection on the intention to continue using the device. The mediating role of trust in the relationship between consumer attitudes and intention to continue using the device is also examined. This study is framed by the Privacy Calculus Theory and the Personalisation-Privacy Paradox (Culnan and Armstrong, 1999; Dinev and Hart, 2006; Xu et al., 2011). It is explored whether the value of receiving personalised information can determine attitudes towards two different types of information collection by the service provider and whether it can also create a trust relationship between users and providers. The results show that the value of personalisation of messages improves users' attitudes towards both types of information collection (overt and covert). This study demonstrates the important role of trust in information management as a positive relationship is found between the value of personalisation and trust, as well as between trust and the intention to continue to use it.
The second empirical study examines whether the humanisation of smart speakers can improve users' attitudes towards covert information collection and build trust. Furthermore, it studies whether such humanisation can influence the reduction of the perception of surveillance exerted by these devices, focusing on the advantages they can exert in terms of social presence. To address these objectives, a theoretical framework is developed based on the literature on humanisation and specifically the three theories mentioned above (Realism Maximisation Theory, Uncanny Valley Theory and Parasocial Relationship Theory). The results reveal that perceived humanisation helps to increase feelings of social presence and also improves users' attitudes towards covert information gathering. It is also shown that humanisation negatively influences trust when humanisation is very low. However, there comes a point of humanisation beyond which its influence is positive. Furthermore, it is shown that humanisation has a negative impact on the perception of vigilance. Finally, this research concludes that if the user trusts the company offering the service, his or her attitude towards the collection of information will be positive, even though users may think that the information has been collected covertly.
The third empirical study analyses how notice and choice options, as tools of corporate privacy policy, can enhance trust in smart home speaker providers. In addition, it examines the role of two types of attitudes (the degree of sensitivity to personal information and the importance of information transparency) on the effect of notice and choice on the effectiveness of privacy policies and on trust. This study is based on the Privacy-Trust-Behavioural Intenttion model proposed by Liu (2005). The results confirm the Privacy-Trust-Behavioural Intent model, which suggests that notice has a direct and indirect effect on trust, mediated by privacy policy effectiveness. On the other hand, choice options have only a total indirect effect on trust. This research also confirms that the privacy policy of smart home speakers can be valued differently depending on users' characteristics or attitudes. In particular, it demonstrates the positive moderating effect of sensitivity to personal information and information transparency on the impact of notice and choice options on privacy policy effectiveness and trust.
This dissertation has several theoretical contribution. It is examined how to eliminate or mitigate two main issues of concern for users: covert information collection and perceived surveillance. It provides new evidence on how to improve user attitudes towards covert information collection and contributes to specific theories related to privacy research. We also contribute to specific theories related to privacy research. In particular, our results contribute to the Personalisation-Privacy Paradox and the Privacy-Trust-Behavioural Intention model. Finally, we contribute to existing literatura on humanisation. This thesis shows that devices with human-like characteristics can improve attitudes towards covert information collection by increasing trust in the provider, creating a greater sense of social presence, and reducing the perception of surveillance.
From a managerial perspective, companies should offer users a personalised experience, as the value of receiving personalised information has been shown to generate greater continuity of use and, more importantly, improve user attitudes. In addition, providers of these devices and services must develop robust and secure transparency, privacy and security policies that assure users that the company is making appropriate use of their personal data, regardless of how it is collected. When it comes to smart home speaker design, attention should be paid to providing them with human-like features, such as more natural language or more human-like interaction.
Abstract (other lang.):
Pal. clave: marketing ; comportamiento del consumidor ; inteligencia artificial
Titulación: Programa de Doctorado en Economía y Gestión de las Organizaciones
Plan(es): Plan 503
Knowledge area: Ciencias Sociales y jurídicas
Nota: Presentado: 25 04 2024
Nota: Tesis-Univ. Zaragoza, , 2024
Contribution of the TFG/M to Sustainability: Fomentar el crecimiento económico sostenido, inclusivo y sostenible, el empleo pleno y productivo, y el trabajo decente para todos. Desarrollar infraestructuras resilientes, promover la industrialización inclusiva y sostenible, y fomentar la innovación. Garantizar las pautas de consumo y de producción sostenibles.
Permalink:
Visitas y descargas
Record created 2024-07-05, last modified 2024-07-09
