Bounded Religious Automation at Work: Communicating Human Authority in Artificial Intelligence Networks

Priestly Authority and the Influence of Power Relations in AI Development

Recent innovations in social robotics and AI have heralded commentaries on workless futures in the so-called second machine age; where electronic substitutes usurp not only manual labor but also cognitive and service-oriented professions once considered immune to automation (Brynjolfsson & McAfee, 2014). Although it has been proposed that work in the service and personal care sectors like clergy and religious personnel are deemed to be among the lowest risks for displacement (Frey & Osborne, 2013), how aspects of this work is changing is less understood. Meanwhile, pastoral work has been envisaged as eliminable labor. For instance, in the article Reverend Robot, the clergy profession comprising of preaching and pastoral care responsibilities is posited to be ripe for systematic and eventual replacement by the technical supremacy of AI agents (Young, 2019). According to ordained clergy and computer scientist William Young (2019), the displacement of human preachers is charted by way of a linear evolution as “[i]t is easy to imagine a progression of steps that lead from the current set of tools to fully automated sermon composition” (p. 492). The four-step work displacement process is described as sequentially unfolding in the following manner:

“(1) Currently available spelling, grammar, and syntax checkers continue to evolve. (2) Automated “research assistants” become available to assemble materials from the web on a pastor-supplied topic or sermon outline. (3) Narrative planning systems become capable of producing a sermon plan or outline to be filled in by a human. (4) Finally, AI systems produce fully developed sermons.” (p. 492)

At the same time, media coverage in a global rush to cover the latest AI often stress the novelty and technical prowess of social robots, together with its transformative effects for religious traditions and hierarchy. Breaking news headlines related to spiritual humanoid robots for example have read “Robot priests, A.I. gods transforming the world of worship” (Chumley, 2019), “AI robot priests replace human clerics with the promise of ‘unlimited wisdom’ (Moran, 2019), “Japan's 'Buddhist' robots replace priests at cut-price funerals” (Billington, 2017) and “Robots are coming for priests’ jobs, too” (Gershgorn, 2017). By emphasizing the attributes of machine learning and the technical accomplishments of AI, the essential motif proposed is the impending automation of religious clergy work. Eye-catching narratives which valorize an AI revolution echo the enduring widespread theme of human annihilation by robots (Richardson, 2015), thereby minimizing the agency of the priesthood and the accomplishments of spiritual workers and volunteers.

Consequently, the spotlight on the novel autonomy of spiritual robotics and its endogenous potential to upend human work raises significant questions for critical communication research surrounding the cultural contexts of robotic development, and their interaction as “independent communicative interlocutors in society” with the rise in human-machine communication (Guzman & Lewis, 2020). According to Fleming (2019)’s critical review, “AI’s potential to replace jobs is too easily equated with its empirical realization….[m]any of the dramatic discussions about robotics and digitalization…often jump from the self-contained achievability of certain innovations (e.g. BlessU-2 giving benediction) to their broader, systematic organizational use” (p. 27). Moreover, as historian, engineer, and robotics pioneer Mindell (2015) observed, “[a]utonomy- the dream that robots will one day act as fully independent agents- remains a source of inspiration, innovation and concern” yet the “precise forms of these technologies are far from certain, much less their social, psychological and cognitive implications” on our traditional roles and ways of living (p. 5).Hence, in contraire to extreme scenarios presented in popular narratives, there is much indeterminacy in AI spiritual developments for human and organizational work that needs to be better conceptualized. In particular, this paper spotlights the role of human influence and traditional authorities as clerics communicate to sustain their legitimacy and ministry power in the evolving AI age.

Historical research on the relationships between new media and religious authority have shown that similar to presently projected AI revolutions, the first flush of digital media (or Internet 1.0 & 2.0) was undergirded by disruptive predictions to religious hierarchy and sites. Initial focus on online phenomena and its disembodied customs together with the potential democratization of information was associated with beliefs that “anybody” can “claim religious authority” (Turner, 2007) even as new mediators like webmasters, online ritual providers and forum moderators facilitated new practices in lieu of cleric oversight (Busch, 2011). However, research studies have illustrated how new technologies are not necessarily critical or damaging to leadership in traditional religious institutions. Instead of viewing new technology uniformly as a threat to religious faith, pastors and priests regarded online platforms of communication as valuable, even crucial tools for connecting to their increasing mediated communities (e.g. Fischer-Nielsen, 2012; Lee, 2009). Diverging from the tenets of secularization theory, clergy from five world religions have expressed how technological modernization and religion is complementary as they adopted new digital technologies to advance their contemporary mission and growth (Kluver & Cheong, 2007). Even digital applications that have been developed to replicate aspects of religious rituals and bypass priests like the Catholic confession app, had acquired approval from traditional papal authority (Cheong & Ess, 2012). Hence though earlier research on the emergence of the Internet was accompanied by logics of disjuncture and displacement of traditional authorities, more recent research studies have highlighted logics of continuity and complementarity while also acknowledging tensions in new media adoption by religious collectivities (Cheong, 2017).

Moreover, historically in several non-Western contexts, emerging technologies including social robots are not characterized by a theme of human destruction but are viewed positively as helpful and harmonious aids to the priesthood and society (Geraci, 2013; Šabanović, 2014). Thus, while traditional clergy has been characterized as potentially having a contentious relationship with the development of newer technologies, authority can be maintained or redressed, constituted by new discursive practices across communication platforms, in the face of AI. To understand how this unfolds, the next section highlights how authority is constructed by ongoing communication practices within religious networks.

Rethinking Human Authority in AI Development: Contextualizing Communication, Continuity and Change

Many robocists have historically applied a techno-centric approach to AI, thereby furthering a largely stagnant view of human capabilities, with scarce consideration of human values in AI development (Markoff, 2016). Regarding human authority, a techno-centric perspective also negates the historical tendencies of powerful leaders to strategically affect technological diffusion and new media interactions in society (e.g. Beatty & Gordon, 1991; Lin, 2003; Siebel, 2019). Processes of technological change are ineluctably associated with a ‘crisis of authority’ given key beliefs in human authority to be an endowed appointment or position held by elites, derived from ideal types rooted in tradition (one’s natural or given place in a system), charisma (personal trait possessed by gifted individuals), and rational-legality (bureaucratic and legal appointments within a hierarchy) (Weber, 1978). Yet as noted by multidisciplinary scholars, this theorization of authority is primarily static, with less attention paid to human communication and the processes of how interaction and consent are constructed among leaders and followers (e.g. Benoit-Barné & Fox, 2017; Lincoln, 1994).

An alternative and more dynamic approach to authority, applicable to deepening understanding of AI development and interaction in society is proposed. Authority can be understood as an order and quality of communication, which is co-constituted between those who acknowledge the asymmetric and consequential nature of their exchanges (Lincoln, 1994). For example, one can observe how human authority works when it is expressed through discourse that not merely describes or reports but impels and establishes precedence or hierarchy (Taylor, 2011). This approach highlights the performative and dynamic aspect of human leadership performances where roles and status distinctions are contested and sustained in ongoing negotiations (Barge & Fairhurst, 2008).

Accordingly, the analytic focus of authority shifts to the continuous process of authoring of agents and practices, which enable certain parties to gain legitimacy and a collective purpose within social units to be coherently composed (Taylor & Van Every, 2014). In the face of new digital technologies, religious leaders have socially shaped user norms (Campbell, 2010) and constructed strategic visions as they negotiated new platforms of persuasion online to achieve robust informational and interpersonal outcomes (Barzilai-Nahon & Barzilai, 2005, Cheong et al., 2011). Given that many AI developments are relatively new and evolving, strategic narratives communicated by leaders about change, the purpose, and the value of AI are important in shaping ubiquitous technology reception and growth in distinctive collectivities and institutional contexts (Boyd & Holton, 2018; DiMaggio et al., 2001; Dourish & Bell, 2011). Specifically, the proclamation of symbiotic AI realities may serve as a form of consequential religious address, as speech acts comes into force for recipients such that the proclaimed word is of higher significance (Løgstrup et al., 2019). In this way, in line with actor-network theory, pastoral authorities may serve as influential translator-spokespersons during periods of technological transitions (Latour, 2005). Clergy can attempt to enroll and mobilize other religious actors by defining their work and attributing specific roles to them, in turn building critical value-laden associations that are aligned to collective or moral goals (Sayes, 2014).

Furthermore, a constructivist communicative approach implies that authority may be distributed or mediated in the sense that sources of authority need not be physically present for authority to be established (Benoit-Barné & Cooren, 2009; Fairhurst & Cooren, 2009). Human authority can be accomplished across time and space, through processes of presentification and invocation, which work to make sources of authority present in everyday talk and ongoing interactions between organizational members (Benoit-Barné & Cooren, 2009; Bourgoin et al., 2020). In the religious realm for example, local and transnational religious workers have constructed the authority of a Dharma master by their in-person and virtual communication that acknowledged her influence over her followers in the global Sangha and branch operations (Cheong et al., 2014). Religious adherents enacted the vision of their spiritual leader and drew insights from her teachings by invoking her sermons and aphorisms to validate the redemptive purpose of their work in their daily lives and on digital media (Brummans et al., 2013). Thus, analytically speaking, the notion of invocation highlights ways in which a principal authority can be made present by the performances of multiple agents, including as argued in this paper, religious social robots. Consequently, a focus on human communication practices can illustrate how authority can work sans physical embodiment, with implications for how clergy authority can be made present in robot-human interactions, in the case of social robot operations and their work.

Third, a dynamic and communication-centric approach to authority brings attention to ways that human leaders can construct social robots as a brand to extend their organizational presence and impact. Branding is at heart, a strategic communication process whereby leaders use narrative techniques to make ideological points, including the crafting of distinctive personalities in the religious arena (Cheong, 2016; Cooke, 2008), to distribute their services and generate cultural capital (Twitchell, 2007). Clerics have exercised strategic communication across the latest digital and social media platforms to create and maintain a coherent image of themselves, even fashioning personal brands as micro-celebrities (Cheong, 2016; Cooke, 2008; Einstein, 2007). Clergy decisions have also impacted ways in which churches and temples consciously design their web presence and virtual platforms to build a consistent and widely recognizable brand identity in order to attract new interactants and advance their outreach (Hoover, 2016; Moberg, 2018; Porcu, 2014). 

A bounded automation perspective stresses how the networks and ambitions of power actors that craft technological development are based on calculated choices and vested interests to control labor, not simply to replace them with intelligent machines (Fleming, 2019). Therefore, as part of human authority practices, social robotics could be co-opted as anthropomorphic organizational mascots to build brand attraction, trust and engagement between religious congregants and spiritual seekers. Correspondingly, the communicative construction of authority and discursive strategies of human leaders have to be accounted for to deepen understanding of the evolution and adoption of AI. To illustrate this, the next section discusses a case whereby key communication practices helped to accomplish religious authority while constituting social robotics.

Religious Robotics: Communicating Authority in AI Networks

In recent years, a robot monk was developed in the Longquan monastery at the outskirts of Beijing, as a collaboration between temple priests and volunteers, and technology and AI companies, and unveiled during the 2015 National Day Gala. The appearance of the robot is based on the character of Xian’er (XE), a novice monk protagonist created for a cartoon series, produced by the temple (Ke, 2016). The robot can explain Buddhist tenets, chant mantras, sense its environment, move via voice command, and hold conversations. XE can also be accessed via its chatbot presence on Weixin or WeChat. Since its initial launch, the robot has undergone several updates, including the miniaturization of its stature, and the extension of its language capabilities to converse not just in Mandarin but also in English. In 2018, the temple signed a deal with China’s leading technology firms iFlytek and Tencent to develop the third generation of Xian’er with “up-to-date AI technology” with improved artificial language ability and augmented data processing functions (Xinhua, 2018). For the purposes of this article, the following discussion focuses on key communication practices that constituted clerical authority in the case of XE in its first years of development (2015-2018), namely preaching and proclamation, invocation and branding.

First, multiple channels of preaching and proclamation constituted the discursive envisioning of the robot not as a priest replacement but as an assistant host. It is notable in its early years how XE was communicatively constituted by the then chief abbot and President of the Buddhist Association of China, as a spiritual guide in adherence to humanistic Buddhist ideals. The Venerable Master Xuecheng’s teachings on digital and social media have repetitively preached the philosophy of Humanistic Buddhism, namely that Buddhism should engage secular society to improve material conditions and alleviate spiritual poverty (Ip, 2009). Accordingly, technological innovations were framed as another pathway of propagating enlightenment and social transformation. Thus, since the reopening of the temple in 2005, this organization’s religious and charity work has been conducted with the use of digital technology (Xinhua, 2015), including computers, iPads, webcams, finger print scanners, and optical character recognition in the digitalization and archiving of Buddhist texts.

Here it is noteworthy how clerics discursively aligned the development of social robotics with scientific knowledge, rationality and modernity. Unlike strands of religious philosophy that stress strict dichotomies between spirituality and science, the physical and the sacred, XE was envisioned and promoted to represent spiritual transcendence. As expressed by Master Xuecheng, in the preface of A Roaming AI Xian’er, the robot “is an attempt which is made by Buddhist practitioners and scientific researchers together, to seek for the truth of our lives….We have the ability to go beyond duality, the conflicts and the contradictions between the spiritual world and the physical world (as quoted in Ke, 2016, p. 14). A digital communication enthusiast, Master Xuecheng has also asserted that the robot monk is the product of the Monastery’s partnership with AI experts for Dharma transmission “so that traditional culture can infuse everyday life, and be reborn in the present moment” to “guide the upcoming era of artificial intelligence onto a healthy road that leads to spiritual insight” (Cheng, 2017).

Accordingly, the lead abbot’s teachings stressed that it is a religious duty for temple monks and laity to incorporate the latest technology including AI into their religious practice, a moral standard not lost on Master Xianqing, a Ph.D. holder in engineering and temple monk for ten years. He commented on his devotion to shape AI development and said, “[j]ust as humans slowly traverse the stages of consciousness, technology has taken the first tiny steps in a long journey toward higher kingdoms of science. We all have a responsibility to guide this journey” (Greenberg, 2016). Similarly, disciple Xian Fan, the head of the monastery’s comic and animation center, stressed that the application of AI was “to spread Buddhist teachings”, so no commercial plans were made for the franchising of the robot monk (Tatlow, 2016). To fulfill the spiritual mandate set out by Master Xuecheng, many lay volunteers including young adults have worked to improve XE’s development and animation, seeing their efforts as a “way of practicing Buddhism” (Ke, 2016). As such, it is evident that proclamations about AI significance by a leading authority functioned as motivating appeals to followers who were consistently positioned as partners with XE to meet global and spiritual challenges with constructive collaboration.

Moreover, it is worth recognizing how in this case, the abbot had preached spiritual compatibility such that “[B]uddhism should keep up with the times and embrace modern technology to promote Buddha’s teachings in an innovative and recipient friendly way” (Tang, 2014). In this light, the discursive envisioning of AI adoption for spiritual enlightenment was both timely and compelling as “Chinese people gain material wealth”, they seek to fulfill new needs for an improved quality of life, and have been described to “have a greater desire for inner enlightenment” (Xinhua, 2016). In addition, the popularity of the temple’s in-house IT Dhyana (mediation) camp among hundreds of employees of Chinese Internet companies illustrated strong interests by faith seekers and followers to harmonize their technology work and spirituality (Lewis, 2016). Thus, by discursively linking AI adoption and spiritual purification, an agenda for social robotics development was set, while preserving cleric authority and the relevancy of Buddhism in contemporary times. In this sense, it is fitting that though XE has been labelled as a robot monk (“ji qi seng”), it is also referenced by other monks as an assistant (“jushou”) and host to guide and engage temple visitors in order to fulfill the temple’s longstanding and religious goals.

Second, communication processes of invocation were operant where the chief temple authority was made present and frequently referenced in human-machine interactions. XE is accessible as a chatbot via its WeChat social media account. In the monastery, visitors can talk to the robot via voice recognition or use a touch-screen tablet display on its tummy. The robot monk can explain basic faith doctrines drawing from data created from Buddhist books, the Abbot’s sermons, and questions and answers on the Abbot’s blog, one of the first blogs created by a religious leader in China, authored daily by him over a decade (Zhang, 2015). In this way, interactions with this robotic agent facilitated the invocation of sacred texts and this particular priest’s teachings. Although XE is voiced by a young Chinese boy, the robot is quite figuratively speaking in the Master’s name, using his discourse. Hence, by invoking the temple chief, the robot monk was not constructed as a destabilizing force for this religious organization but as countering a sense of deterritorization made possible by mediated interactions; making present cleric authority through its emergent talk with temple visitors.

Specifically, cleric authority was made present in human-machine communication through supplicants’ questions and statements. As Kevin Kelly, a Christian co-founder of Wired magazine who argued for a need to develop “a catechism for robots” noted, AI systems are biased like “the clergy is bounded too”. Although everyday users do not know how AI black boxes are coded, they are nonetheless subject to the interpretative work of AI expert systems when religious robots offer answers to spiritual interactants, “[e]ven if it doesn’t presume to interpret the verse for you, in choosing that verse it’s already doing hidden interpretational work. It’s analyzing your situation and algorithmically determining a recommendation” (Samuel, 2020). In the case of XE, it is particularly striking how cleric authority is reinforced during in-person and virtual question and answer sessions. If XE does not understand a question posed, it will say, “Wait, I will ask my Master” or “I need to check with my Master (Shifu)” (Tatlow, 2016). These responses veritably elevated the chief priest as the more credible spiritual authority.

Moreover, it was observed how monks and volunteers drew upon insights derived from their leader’s teachings and invoked his words to complete the machine learning work that sustains XE’s daily operations. Consider volunteer Luo’s experience as he described his work to help process answers to questions posed to XE; he said “[a]lthough I was lucky enough to listen to the Masters’ instructions, I still had to refer to these instructions when answering questions. During that time, I am always very delighted. It might be that the speeches given by the Masters make me more respectful, or they add more wisdom to my life, which I will never acquire during my entire life.” (Luo, 2018) Consequently, as spiritual workers and volunteers cited and incorporated their leader’s instructions in their work to give meaning to their AI experiences, invocational practices played a vital communicative role in the co-enactment of this religious organization, and the ongoing construction of its social robotic developments.

Third, it is interesting to observe how this robot monk was simultaneously constructed as a brand for prominent media and temple outreach. Besides digitalization work, the process of “making” this robot also occurred through everyday material and mediated communication including the production and circulation of organizational symbols and the celebration of the robot monk in prominent temple and public events. Here, although commercialization of XE and XE paraphernalia was discouraged (Ke, 2016), publicly showcasing the robot monk was paramount to this temple’s religious mission of promoting the relevancy of contemporary Buddhism as the abbot explained, “with the image of Xian’er, people can easily understand monks, monasteries and Buddhism” (Han et al., 2018).

Related to communication then is the iconic (re)production of XE as a unique cultural symbol of the temple. XE images are pervasive online and offline, including on temple grounds, mounted on wooden signs that literally and figuratively point temple visitors in the right direction to different key locations. In the animation center where the robot is typically housed, the temple had provided XE with its own name card (stating its temple address, telephone number, and QR matrix barcode to its WeChat social media handle) and customized large floor stand-up poster boards that also featured XE’s images and its QR code. In this regard, news reports have highlighted how XE’s image and interactivity functioned as a major attraction to draw visitors to the temple. It was reported for example how Liu, a non-Buddhist who visited the temple to meet and pose for pictures with XE said, “I feel it is like a temple mascot, making Buddhism much more accessible” (Lu & Robertson, 2016).

Furthermore, as XE was regularly exhibited in temple and AI events around the country, the work of branding AI here was accomplished by human gatekeeping since access to the robot was strategically managed to present a consistent organizational front. Within the temple, XE was carefully watched by staff as they chaperoned him in the animation center or at public events like the international Children’s animation festival. XE was also accompanied by an entourage of monks and volunteers in AI & robotics and Buddhist conferences around the country. In this aspect, it is critical to point out that contrary to the hopes and predictions of robot autonomy, this robot’s day-to-day operations appeared firmly embedded in networks of human authorities, support and care.

Conclusion

This paper sought to provide a critical perspective to human-machine communication through a three-fold assessment of AI powered religious robotic developments to advance understanding of bounded religious automation, in light of human authority practices and power relations within spiritual organizations. By proposing to understand cleric authority not solely as fixed appointments but as dynamically constituted by communication, the contention of future work and job displacement propositions are contextualized within every day discursive norms that involve both strategic ‘top-down’ and ‘bottom up’ invocational practices in human-machine communication enacted within the spiritual domain.

In the face of rising AI development in increasingly mediated societies, traditional authorities can pragmatically adopt new technologies and preach the value of AI as reinforcing the legitimacy of their spiritual work, thereby imputing a religious mandate to develop and upgrade AI. Specifically, the case of the robot monk illustrated multiple ways in which a prominent cleric had framed AI in a manner that not only allowed social robotics a place in religious practice but that also demanded the integration of AI in his organizational practices and temple outreach. In this way, rather than AI presenting an inevitable clash between spirituality and science, the status quo in this case was maintained and cleric authority was simultaneously accomplished through related invocation and branding activities.

Therefore, while projected claims about social robotics underscore their potential to unseat human experts and elites to dramatically disrupt the future of work, AI and society debates must continue to examine how human agency and authority can be redressed and constituted by strategic communication practices. Here, it was significant to discuss how recurrent and multiplatform practices of preaching, invocation and branding in the recent initial years of XE development have built positive associations between spiritual values and new robotic developments to sustain historical hierarchies and relational bonds. In light of anticipated changes with AI advancements including technological “singularity”, future studies in social robotics can help develop a more comprehensive and acute understanding of power relations in human-machine communication. Empirical examinations of AI developments in specific organizational and ideological contexts can help counter pervasive cultural narratives which assume a flat autonomy between human and non-human agents in AI networks.

Specifically, the discussion here has broader implications for critical communication research on the intersection between religion, power, and human-machine communication. The case of Xian’er highlights the salience of inspirational leadership in the success of AI adoption in particular, and organizing for spiritual change more generally. As this study shows, followers and volunteers motivated themselves to adopt the latest mediated AI technologies by invoking a prominent leader in their day-to-day interactions, thereby sustaining traditional religious authority. While case illustrations here focused on the pioneering phase of XE’s development, further long-term research can delve into the power of organizational leaders on AI cooptation, beyond the limits of their term in office or official appointment. On this account, it is noted that the chief abbot of the monastery stepped down in 2018 after a report filing sexual harassment and corruption charges against him was posted on social media. While the incident has been characterized as part of the Chinese #metoo movement (Barron, 2018), it nonetheless illustrates dynamism in leadership performances, as ongoing communicative practices (and its lack thereof) affect the accomplishment of authority. Interestingly, the influence of the chief abbot on XE’s development has purportedly endured to a certain extent, as Travagnin (2020) recently reported that even after Master Xuecheng’s downfall, his teachings have been translated to XE’s cartoons to support the continued popularity of XE the robot monk. Thus, future longitudinal research should track the continued practices and ways of enacting religious authority with implications for AI development over time.

Furthermore, observations made here about XE and elsewhere about the state of spiritual robotics illustrate not only the current communicative limitations of robotic priests in terms of conversational fluency (Cheong, 2020; Travagnin, 2020) but also underscore the complex regulatory networks in which spiritual machines are embedded. While AI developments are ongoing and the future is inherently unknowable, the robot monk at the point of analysis was not yet autonomous or independent. Though Xian’er has human likeness, its appearance (less than half adult height), utterances (limited vocabulary), and motions (difficulty with traversing uneven grounds), mark it distinctly as a robotic agent, and the constrains of its mechanical form have been detected. For example, according to Wang, a Buddhist and temple volunteer, “I don’t think it’s possible for the intelligence of a robot to be advanced enough to understand human feelings…The robot doesn’t really live up to my expectations.” Another temple visitor Zhang commented, “It relies on permutations and combinations of words to solve problems, but whether it can really deal with deep personal issues, I’m not sure” (Tatlow, 2016). Recent ethnographic observations of XE’s human-machine communication have also highlighted specific communication affordances and constraints; the latter included errors in its voice search function and natural language processing, limited facial and object recognition, limited mobility and battery capacity (Cheong, 2020). Hence, technology singularity is far from imminent and concerns expressed over clergy displacement appears unwarranted at this point.

However, it is recognized that given ongoing developments and potential advances in AI and computing graphics, synthetic media or deep fakes may be constructed in the virtual religious domain to deceive or mislead faith adherents. Religious chatbots can be hijacked and alternative personalities mimicking religious leadership can present fake content on social media platforms. As such, future research could probe into new practices that religious leaders construct to dynamically communicate their credibility and authenticity, including policies to counter misinformation shared on their platforms and through robotic agents. Future research should also explore how and to what extent religious congregants and seekers classify or interpret human-machine communication as genuine, efficacious or helpful to fulfill their spiritual needs. In particular, many religious organizations like the monastery in this case, have established media units or multimedia ministries. Thus, it would be interesting to examine how the work of these staff and volunteers help to (re)construct clerical authority, for instance by performing the coding, translation and chaperone duties as outlined in this case, or in other ways like content moderation or censorship of mediated interactions. In addition, while it was beyond the scope of this paper to trace how all the related operations of XE comics and chatbot were managed by temple staff, future research can investigate practices of media governance to explicate the control (or lack thereof) of XE’s mechanical representations and interactions in light of contemporary convergence culture.

Last but not least, future research can examine in more in-depth ways, other cases of social and spiritual robots as they emerge. In the long term, examining human communication practices can provide insights into the ways that leadership, culture and AI intersect to impact participation and power, to enlighten narratives on a post-humanist future debating whether human subjects are decentered or still germane to society (Bolter, 2016). The continued development of critical approaches to AI helps us to advance understanding of how authority practices in organizations discursively restructure work to manage AI innovations in line with historical practices and collective values, to avoid the fetishization of ‘smart machines’ or the treatment of the latest robots in social vacuums. Through refocusing the question of who are the new communication interlocutors in human-machine communication, to what conditions they work and how the authority of their human leaders is constituted, it is hoped that new theoretical and practical interventions to incorporate robotic agents are generated to enable human flourishing and virtuous growth.