Recent advances in AI robots’ use have gained increasing attention across a variety of academic disciplines, including service research. Nonetheless, there is little consensus on the ethical implications of AI robots’ application to services; and the debate is primarily at the dichotomous level of ‘unethical or ethical’. To address the need to build an ethical evaluation that goes beyond the unethical/ethical binary, this study creates, explicates, and offers evidence for a new conceptual framework that explains the ethical implications of AI robots’ use. The new framework builds on normative business ethics. It differentiates between four layers of moral intensity: (a) illegal, (b) immoral, (c) permissible, and (d) supererogatory pertaining to the application of AI robots. We combine the moral intensity element with the consideration of the locus of moral accountability across the continuum of fully autonomous to fully assisting AI robots. AI robots are defined as intelligent machines, software and systems that possess the increasing ability to learn and arrive at decisions. They can increase efficiency and increase quality of life (Čaić et al., 2018), for instance by assisting in elderly care (Broekens et al., 2009; Jiang & Cameron, 2020), support medical diagnoses (Yoon & Lee, 2019), and ease mobility in the case of autonomous cars (Hassan et al., 2018). Concurrently, using AI robots in services prompts societal changes and thereby yields ethical implications (Wirtz et al., 2018). Efforts devoted to relevant research ethics are necessary in order to – at least partially – alleviate relevant risks (Russell et al., 2015). Normative business ethics is beneficial for this study due to its link to the ‘real world’ that derives from its focus to critique practice by reference to the ideal (Trevino & Weaver, 1994) and to discourage firms from taking unethical gains, even in cases where legal directives are not feasible (Heath, 2014). Normative business ethics concentrates on how actionable responses can be formulated in response to ethical challenges, instead of the explanatory discussion of ethically significant issues in descriptive business ethics. Applying these lens enables to identify obligations of consumers, companies, and the government in using AI robots in services by incorporating both conceptual efforts as well as implications for practice (Hasnas, 1998). We link the four categories of illegal, immoral, permissible, and supererogatory into the conversation about moral intensity. Moral intensity is defined as the extent of issue-related moral imperative across different circumstances and considers the impact an individual action can have on multiple stakeholders (Jones, 1991). Nonetheless, actions emerge from a web of connections and thus it is relevant to assess the extent to which an individual or firm is accountable for outcomes – thus, the notion of the locus of responsibility has been established to assess this (Nicholson & Kurucz, 2019). The study opens the academic conversation towards moral accountability between AI and humans, taking into account the moral intensity of service use cases. Ethical implications are discussed for regulators, a future research agenda is provided on the ethical implications of AI robots’ use in services.

Servicescapes and Augmented Reality in Heritage Tourism: Heritage Producers Perspectives

Purpose: The purpose of this research is to explore the incorporation of augmented reality (AR) in the designing of servicescapes at heritage settings, utilising heritage producers’ insights. Servicescapes, specifically in heritage tourism, are recognised as staged spaces that can direct consumer reactions (Chronis, 2019), raising the question of how “augmented” servicescapes are formed (Amitrano et al., 2021).

Methodology / Approach: In-depth qualitative interviews are being conducted with experts situated in heritage institutions across the United Kingdom who have been involved in developing experiences that include AR. Data collection focuses on understanding the manner in which AR is utilised and is expected to add value to the visitor’s experience at eight heritage settings.

Findings: Preliminary findings indicate that virtual and physical elements are symbiotic in the “augmented” servicescape. Virtual elements are designed to function in conjunction with physical elements in situ. Additionally, the social element is identified as an accompanying component, with interactions amongst visitors and service providers encouraged and considered vital when designing a servicescape that incorporates AR. It is intended that consumers explore personal interpretations of physical and virtual elements, supported by social aspects, ultimately co-creating the “augmented” servicescape, a suggestion that is supported by previous conceptualisations of servicescapes (Ballantyne & Nilsson, 2017).

Research Limitations: This study explores the supply side of the heritage experience, identifying the methods and strategies employed when designing a servicescape, utilising experts’ insights. Understanding consumers responses to the “augmented” servicescape by conducting ethnographic research would further complement knowledge of effectively utilising virtual technologies in service experiences.

Practical Implications: With virtual technologies continuing to be novel in the heritage sector, managers need to be aware of employable strategies that create effective “augmented” servicescapes. Findings from this study provide recommendations for managers regarding how to utilise the virtual elements offered by AR in a manner that initiates consumer interaction with the offerings, and ultimately engagement.

Originality / Value: Current research of digital servicescapes has maintained focus on network based services, such as social networks, with newer technologies that integrate digital information into solid environments overlooked (Schmitt, 2019). This study sheds light into how one such recent technology, AR, influences the way servicescapes are being designed, utilising the context of heritage tourism that has previously received lesser attention in the servicescape context.

References:

Amitrano, C. C., Russo Spena, T., & Bifulco, F. (2021). Augmented Servicescape: Integrating Physical and Digital Reality (Digital Transformation in the Cultural Heritage Sector (pp. 181-197). Springer.

Ballantyne, D., & Nilsson, E. (2017). All that is solid melts into air: the servicescape in digital service space. Journal of Services Marketing, 31(3), 226 –235.

Chronis, A. (2019). The staging of contested servicescapes. Journal of Service Research, 22(4), 456-473.

Schmitt, B. (2019). From atoms to bits and back: A research curation on digital technology and agenda for future research. Journal of Consumer Research, 46(4), 825-832.

Retailers increasingly implement artificial intelligence-based (AI) technologies to collect high quality consumer data in order to create personalized product offerings and reduce return shipping (Aiello et al. 2020; Arnett 2019). Puntoni et al. (2021) highlight that consumers do not only feel served by the rather effortless AI-based service (Ameen et al. 2021), but also feel observed as their control over personal data is vastly diminished. This specific loss of control over personal data differs from what privacy literature has investigated so far. Whereas AI-based data disclosure processes are similar to prior approaches that are actively initiated by consumers to fulfill a distinct purpose (Krafft et al., 2021), they differ substantially as the AI-based service then takes over the actual task of collecting and processing personal data on their behalf. Thus, we need to understand how consumers perceive their control over personal data in this new data disclosure setting and which consequences this has on their intention to enter AI-based data disclosure processes.

Prior research cannot answer these questions as existing operationalizations cannot adequately capture consumers’ control perceptions in AI-based data disclosure services. Privacy literature has, on the one hand, treated and measured perceived control as a one-dimensional construct (e.g., Xu et al., 2011), which is too imprecise to capture consumers’ specific loss of control in AI-based data disclosure processes. On the other hand, it has focused on measuring single distinct control dimensions such as access and use control (e.g., Zhao et al., 2012) that are likely not affected in AI-based data disclosure. This is problematic as understanding the nature of how consumers perceive a loss of control over personal data is crucial in order to precisely address it.

To help marketing scholars and managers better understand consumers’ control perceptions in AI-based data disclosure, we pursue the following research goals: first, we aim to extent the control-based privacy literature by developing a multi-dimensional perspective of consumers’ perceived control over personal data. Second, we intend to holistically capture consumers’ control perceptions to better understand which control dimensions are affected in AI-based data disclosure processes. Finally, we aim to develop mechanisms that are capable of further addressing consumers’ specific control perceptions in AI-based data disclosure services such that all parties involved can appreciate the facilitated process and its benefits.

To pursue the identified research objectives, we conducted a structured literature review and found that research has (implicitly) distinguished between four control dimensions (i.e., control over personal data collection, submission, access, and use). We validate our assumption that consumers’ can distinguish those dimensions in study 1 (N=266). Study 2 (N=190) confirms that consumers’ control over personal data collection and process effort are decreased in AI-based data disclosure processes while access and use control perceptions are not; surprisingly, consumers do not perceive less control over personal data submission. The future objective of this research is to investigate mechanisms that increase consumers’ control over personal data collection in AI-based data disclosure settings.