A Lack of Agency: Artificial Intelligence Has So Far Shown Little Potential for Church Innovation—An Exploratory Interview Study with Protestant and Catholic Leaders in Germany

Abstract

This study explores the use of artificial intelligence (AI) in religious leadership in Germany, focusing on the interplay between technological innovation, theological principles, and human interaction. Drawing on qualitative methods, 23 Christian leaders and experts were interviewed to examine their perceptions, assessments, and potential applications of AI and related technologies in their work, alongside ethical and theological considerations. The findings reveal a prevailing ambivalence towards AI: while it is generally accepted as a tool for administrative tasks, its use in pastoral contexts encounters resistance due to ethical concerns and theological tensions. Despite predominantly neutral to positive attitudes, many leaders lack proactive engagement in exploring AI’s transformative potential—pointing to a marked lack of agency. Digital competence among leaders emerges as a significant factor influencing the openness to AI adoption. This study identifies key barriers to the integration of AI into religious practice and underscores the need for strategic education and planning. It advocates for a balanced approach to leveraging AI in ways that align with religious values while embracing innovation in a digitalizing society.

1. Introduction

With the ongoing advancement of artificial intelligence (AI) for everyday use, new horizons are also opening up for its application in the field of religion (cf. Cheong and Campbell 2024, Part I). This study therefore examines how religious leaders understand AI, what forms of use they consider possible within their areas of responsibility, and what future visions they develop for a church engaging with AI. It focuses on leadership figures within the Protestant and Roman Catholic Churches in Germany.

This contribution is situated within the broader international discourse on Digital Religion (cf. Cheong and Campbell 2024) and Digital Theology (cf. Kurlberg et al. 2025), as well as among the emerging body of work that explicitly addresses AI in religious practice (cf. Trotta et al. 2025). Within this discourse, leadership profiles such as the digital entrepreneur, digital spokesperson, and digital strategist have already been conceptualised (cf. Campbell 2020). However, the central concern of this study is not media usage typologies but rather the narratives through which actors describe and interpret (a) their commitment with AI—particularly including the tensions between institutional innovation and tradition—and (b) the options for action they are considering. Agency therefore is understood as meaningful action, which in times of digitalisation is shaped by the perception, selection, interpretation of, and interaction with media content in one’s own life (cf. Eichner 2017; Radde-Antweiler and Nord 2026).

Furthermore, this contribution arises from a trinational research project on AI and religious leadership. The results presented here were obtained in the German context and in the German language and are now being presented in English. Pauline Hope Cheong (Arizona State University) and Ilona Nord (University of Würzburg) initiated the interview study with context-specific approaches and different survey groups in the USA (Cheong and Liu 2025) and Germany in winter 2023. In 2024, Thomas Schlag, Gabriela Frey, and Katharina Yadav joined the project for the Swiss context, with slightly modified questions (cf. Schlag et al. 2025a). This study can also be seen as a follow-up to the project Churches Online in Times of Corona (cf. Schlag and Nord 2023; Schlag et al. 2025b). That study surveyed church professionals about their use of digital communication and their assessment of its significance for their areas of work.

2. Contextual Reflections and Methods

2.1. Contextual Reflections

In April 2020, the international and ecumenical research team CONTOC («Churches Online in Times of Corona») was established (cf. Schlag and Nord 2023) to examine the development of digital communication practices in churches under pandemic conditions. This largely quantitative study, conducted via a questionnaire distributed broadly among the clergy, was followed in spring 2022 by a second CONTOC study. This follow-up, a purely Protestant, binational project between Germany and Switzerland, extended its target group to include all professional roles within the church (cf. Schlag et al. 2025b). Its aim was to explore the use and evaluation of digital communication formats in ecclesiastical contexts, reflecting on the experience of the pandemic period. The findings pointed to the relevance of a qualitative research approach to gain deeper insights into the argumentation patterns of church leaders. Even the evaluation of the open-ended responses in the initial CONTOC study had already proved fruitful in this regard (cf. Yadav 2025).

But it is also necessary to show how this interview study differs from the previous ones in two key respects: first, it is conducted independently of the pandemic context, and second, it focuses on experts who, by their own account, consider themselves to be tech-savvy leaders, particularly in the field of digitalisation and AI within the context of churches and Christian theology.

Secondly this article forms part of the journal issue “Institutional Innovation and Tradition: Understanding Changes and Continuities in Religious Leadership, Communication, and Technology Ethics”. It may seem surprising that this contribution places an emphasis on the polarity between institutional innovation and tradition. Contributions from the German-speaking discourse in practical theology—among others, those drawing on qualitative research into digital church publics—often favour the concept of lived religion over that of the institution or church as the central category for understanding the Christian religion. This perspective aligns with research that explores how individual actors present themselves in religious terms on social media. In the German context, for instance, pastors often understand their online presence as a form of strategic communication on behalf of a religious organisation, which they translate into corresponding communicative practices (cf. Weyel 2025). However, in relation to the Christian vocation of being a public pastor, there is a need not be linked to an institutional understanding: “It is not primarily the institutional context that is associated with professionalism, but rather the mode of credible and convincing self-presentation, which brings its own set of demands. The pastoral-theological question of whether the office carries the person, or the person carries the office, no longer arises in this form” (Weyel and Stolz 2025, p. 10).

In this context, the thesis that the office no longer carries the person becomes comprehensible. Church personnel are now expected to persuade through personal authenticity; they can no longer rely on the authority or reputation of the institution, which has become increasingly fragmented. Conversely, it is also argued that pastors are no longer capable of carrying—or in other words, representing—the office, as the significance of the institution for the Christian mission of communicating the Gospel has, in the wake of secularisation, gradually lost its central role (cf. Swart 2022). This describes the situation in the German context. Practical theologians from the Global South, by contrast, have drawn attention to the limitations of this line of argumentation, emphasising that the role and relevance of the institutional church has not eroded to the same extent in contexts beyond Europe.

On the contrary, the church is currently experiencing renewed acceptance, and the institutional dimension of religion—manifest in the form of churches—is gaining in significance. According to Swart (2022), global migration movements may even prompt processes of reinstitutionalisation in the Global North. In the European context, it is also worth considering whether the digital formats of lived religion, currently under study, still benefit from institutional logics—logics that may no longer operate in traditional forms but nonetheless persist in fragmented or transformed ways. Christian lived religion cannot be entirely disentangled from institutionalised contexts—whether it is practised within them, shaped by their environment, or positioned in contrast to them. This applies not only to religious practice itself but also to its academic investigation and interpretation.

For these reasons, it is essential to seek out institutional innovations in the context of digitalisation and datafication processes—also within traditions that are embedded in the very institutions in need of renewal. This includes examining which traditions—such as specific sources from the history of Christianity—can inform, shape, and profile institutional approaches to technologies like AI. Religious leadership fundamentally relies on an ongoing process of reflecting on innovation in relation to interpreted traditions. This is not an extraordinary challenge for the group under study but rather constitutes the core responsibility of church leaders across all areas of their practice (cf. Preul 1997). The term church institution is understood here as a societal structure in which the reciprocal interpretation of innovation and tradition—constitutive of Christian religion—is cultivated, safeguarded, and advanced. Where such processes take place, churches hold the potential to contribute to society—for example, in the spirit of a public Protestantism or as intermediary institutions—by engaging in the reflection on pressing issues, such as the use of AI, and by contributing to ethical debates or profession-related discourses both within and beyond their own institutional contexts (cf. Huber 2000; Albrecht and Anselm 2017).

Moreover, the use of the terms institution and institutional in this contribution reflects a certain sociological ambiguity. In the fields of Digital Religion and Digital Theology, the distinction between the institution and organisation is often blurred. As such, the term institution is understood here to include organisational dimensions, while also holding together the fundamental meaning-making structures that mediate between tradition and innovation and enable their practical implementation. The operationalisation of the relationship between the individual and institution—including the church as an organisation—was also addressed in previous CONTOC studies. The various professional groups involved expressed a considerable need for further training—understood not merely in technical terms: “It is therefore necessary to ask how the ‘traditionally proven’ can be combined with the ‘experimental and new’ in a meaningful and appropriate way within the horizon of a public church and theology which are life-serving, relational, and sensitive to vulnerability” (Schlag and Nord 2023, p. 535).

In summary, it is the areas of tension between the “person and institution” as well as “innovation and tradition” that this study aims to explore in more detail. For this purpose, it asks for futures and competencies related to AI in church contexts, for applications in religious leadership, and tries to figure out the barriers which prevent experimentations with AI. It finally thematizes theological principles and values as well as personal experiences with AI and digitalisation (see four Hypotheses).

2.2. Methods

Data collection was performed during 15th of January until 17th of March 2024 by conducting video conferences (Zoom) and afterwards transcribing the interviews using the software f4; in a second step, the manuscripts produced were compared and checked against the video recordings. For the analysis, firstly a content-analytical method was chosen (cf. Kuckartz and Rädiker 2024). The data material was primarily coded deductively. The analysis process was supported by the software MAXQDA 2022 (cf. Schneijderberg et al. 2022).

Key questions, as just described above, were previously explored in discussions on results of the CONTOC Studies and specifically findings from Campbell’s typology of digital leadership (cf. Campbell 2020). These results let us expect that analytical categories for coding responses like “innovation readiness,” “relationality,” and “authority delegation” could become important. These categories were supplemented by insights from Digital Theology (Kurlberg et al. 2025), guiding the interpretation of theological reflections on AI.

In addition, the dataset was evaluated using a sentiment analysis. Sentiment analysis is a quantitative automated method of text analysis from the field of Natural Language Processing (cf. Schneijderberg et al. 2022, pp. 278–357). This approach serves the research objectives at hand by enabling an exploration of leaders’ openness to innovation not solely through content patterns of argumentation and their logic but also about the forms of expression associated with these arguments (positive–negative–neutral). Sentiment analysis can thus productively deepen the content analysis, because the topics filtered out in the content analysis can now also be seen in relation to the emotional or affective evaluations they contain.

3. Sampling

Participants were selected through targeted invitations to church leaders in Germany, aiming to capture current competencies and forward-looking positions in dealing with digitalisation and AI. A total of 23 interviews were conducted with leaders from various Protestant and Catholic fields of work, representing a range of professional backgrounds, including individuals with a focus on theology, education, and church public relations. The sampling focuses moreover on individuals whose work is directly related to religious practice and theology in a narrower sense. No church IT specialists were interviewed. The sampling is presented below in

Table 1. An overview of the sampling (captions: No = the number of interviewed persons; M = male; and F = female).

Interview	Job Title	Gender	Denomination
1	Deaconess and referent of inclusive church (r)	F	Protestant
2	Teacher of religious studies (l)	M	Prot.
3	Pastor (l)	M	Prot.
4	Senior church councillor/digitalisation unit (r-n)	F	Prot.
5	Freelance editor for religious institutions (l-r-n)	M	Unknown
6	Scientific assistant in digital religious studies (r-n)	M	Prot.
7	Head of studies at a Protestant academy (r)	M	Prot.
8	Digital education officer at the religious education institute (n)	M	Prot.
9	Head of church service in a diaconate (r)	F	Prot.
10	Head of the media competence office (n)	M	Roman Catholic
11	Religious educator with research centre for public religious education (n)	F	Prot.
12	Teacher religious education/digitalisation officer (l-r)	F	Prot.
13	Self-employed theologian in the field of digitality/church (l-r-n)	M	Rom. Cath.
14	Church representative for digitalisation (n)	M	Prot.
15	Pastor/internet Officer (l-r)	M	Prot.
16	Pastor/speaker for digital church (l-r-n)	M	Prot.
17	Digital department (n)	M	Rom. Cath.
18	Head of communications and media relations for an archdiocese (n)	M	Rom. Cath.
19	Research associate, assistant professor (l)	F	Prot.
20	Diocese communications department (r)	M	Rom. Cath.
21	Research associate (l)	M	Prot.
22	Deputy academy director/head of studies (r)	M	Prot.
23	Head of an institutional department for religion (r)	F	Prot.

in order to provide an overview of the interviewees. It comprises 17 Protestant and 6 Roman Catholic leaders, chosen purposefully for their self-identified involvement with digitalisation as experts.

It should be noted that no data was analysed with regard to religious denomination or gender, but this socio-demographic data should nevertheless be disclosed. The job titles show that three different levels of the range/degree of responsibility were taken into account: local (l), regional (r), and national (n). The work of some participants relates to all three ranges, because they address their work to groups of persons with a local, regional, and national outreach, but it does not include the responsibility for personnel or organisational structures.

In order to preserve the anonymity of the interviewees, the statements were pseudonymized. The evaluation will relate to the interviews by speaking of person 1 to person 23, thereby reflecting the order in which the interviews were conducted.

4. Central Hypotheses

The hypotheses which were developed for this study address the agency of religious leaders by making their commitment explicit (engage, envision, identify, and express) and by asking for options of action as follows:

H1. 

Definition and Understanding of AI: Religious leaders who engage with AI develop distinct definitions and understandings of artificial intelligence that reflect their theological, ethical, and practical concerns, including the perceived potentials, limitations, and risks of AI.

H2. 

Futures and Competencies: Religious leaders who envision future applications of AI in the church and society tend to articulate specific competencies—such as ethical discernment and digital literacy—as essential for responsible leadership in times of technological transformation.

H3. 

Applications in Religious Leadership: Religious leaders identify areas of ministry in which AI can support or enhance their work, while at the same time maintaining that core spiritual and pastoral responsibilities require an irreplaceable human presence and wisdom.

H4. 

Barriers and Human–Machine Distinctions: Those who express reservations or reject AI in church contexts often base their stance on a theological or philosophical distinction between human and machine capabilities, especially concerning a higher authority and pastoral care.

H5. 

Theological Principles and Values: Religious leaders who engage with AI draw on established theological principles, ethical frameworks, and Christian values to evaluate the use and development of AI in alignment with their religious convictions.

H6. 

Personal Experience with AI and Digitalisation: Religious leaders who report more direct and positive experiences with AI and digitalisation are more open to exploring its use in their field of work and tend to evaluate it more constructively.

5. Findings

The following section first presents the findings from the content analysis of the interviews, structured according to the hypotheses outlined above (Section 5.1). This is followed by a presentation of the results of the sentiment analysis (Section 5.2).

5.1. Results from the Content Analysis of the Interviews

• Ad H1: Pragmatic Understanding of AI, Largely Without Direct Impact on Church Contexts

The interviews reveal that there is at least a basic—and in some cases a more advanced—understanding of how AI functions from a technical perspective. However, respondents often limit their descriptions to specific applications, such as language models used for answering questions, writing texts, translating, or programming. The religious leaders interviewed in German churches largely frame AI as a functional tool. Supportive roles and the practical utility of AI are consistently emphasised. Some interviewees demonstrated an awareness that AI exists in various degrees of complexity, including forms that are already used unconsciously. Yet, their potential within church contexts has scarcely been the subject of reflection, such as the distinctions between weak and strong AI. One interviewee, for example, stated “This is not artificial intelligence we are dealing with here, but rather an algorithm-based system with neural networks” (Interview 8, Position 33).

Another respondent noted, “AI cannot think and reflect upon its actions, this machine has no idea what it’s doing” (Interview 10, Position 58). At times, statements were only partially accurate or formulated ambiguously: “Algorithmicity is something we cannot grasp and something that no longer follows [sic] logical paradigms. We must avoid loss of control” (Interview 21, Position 57).

These statements indicate a general understanding that some algorithms are based on formal logic, while others—like machine learning or neural networks—are based on statistics, probabilities, and pattern recognition. Overall, boundaries and risks heavily shape the understanding of AI. This likely explains why few respondents made a direct connection between AI and the church context when asked about their understanding of the technology. AI was not discussed as a “gamechanger” amid church crises nor as a technology fundamentally altering their own work.

Both the interviewees and the institutions they represent do not expect AI to have significant transformative effects on their personal or institutional activities.

• Ad H2: Hesitant Visions of the Future, Orientation Gaps, and Identified Shortcomings

The church and theological actors interviewed display a high degree of realism; artificial intelligence is rarely associated with visionary or future-oriented thinking. Instead, the focus lies on practical advantages—such as the use of administrative software or communication support through large language models. “Of course, it requires a certain openness to actually use and experiment with these tools” (Interview 20, Position 103). These advantages are often related to individual workflows: “We’re already experimenting with it […] for corrections, alternative suggestions, and shifts in perspective. Getting summaries of articles, for example” (Interview 10, Position 70). In areas such as pastoral care and public relations, the potential of AI is generally acknowledged but often limited to specific, clearly defined functions. For example, it is seen as useful for locating individuals or facilitating initial contact. Overall, the use of AI is viewed critically, particularly due to concerns about the lack of personal connection in communication. “You can’t replace AI for a nice evening with your wife over a glass of red wine” (Interview 23, Position 52). Some participants call for a more critical engagement with AI and stronger regulation: “[…] when it comes to AI, we are ethically lagging behind” (Interview 16, Position 36). The use of “we” in this statement refers to the institutional or organisational context and signals that the ethical dimension is seen as crucial in dealing with AI—precisely where, according to the respondent, a coordinated, church-wide commitment is currently lacking.

• Ad H3: Scepticism or Rejection of AI in Pastoral Practice

As previously noted, church leaders in Germany tend to accept the use of AI for administrative and organisational tasks: “I do a lot of media work for the church—various kinds of programmes, websites, and digital media” (Interview 5, Position 2). However, when it comes to pastoral–theological responsibilities, the use of AI meets with a clear resistance. In areas such as pastoral care and core theological duties, AI is regarded as unsuitable. One respondent, speaking about the idea of androids or AI-driven robotics, put it bluntly: “Preaching? No. Pastoral care? No. I have to be honest... I just don’t see a robot folding its hands and praying for me. I believe that’s something reserved for human beings” (Interview 3, Position 40).

Scepticism is also evident regarding interactions with members of the faith community: “The question is whether AI can truly grasp the nuances of interpersonal communication, which only function between human beings” (Interview 7, Position 20). Even if AI were capable of social interaction, the search for meaning and the need for authentic communication are seen as fundamentally human concerns: “The search for meaning is real, and I maintain to say: is it not the human being it therefore requires, also for conversation, for genuine exchange?” (Interview 2, Position 26). In the realm of social interaction, further concerns were raised about AI reinforcing harmful stereotypes:“Traditional gender roles and even fundamentalist religious views are reproduced uncritically by AI—and I do see that as a real danger” (Interview 15, Position 71).

Two central lines of reasoning—grounded both in personal convictions and in what respondents describe as a “Christian attitude”—are cited to justify the rejection of AI in liturgies for special occasions, worship services, and especially in pastoral care: (a) Communication with AI is regarded as an interaction with a non-human agent, which is deemed inappropriate for ecclesial communication. (b) Appropriate, non-discriminatory church communication and meaningful interpersonal contact are seen as possible only through human-to-human interaction. Overall, the patterns of argumentation suggest that scepticism and rejection are primarily shaped by unresolved ethical concerns and the perceived risk of losing control.

• Ad H4: Barriers to AI: Lack of Relationality and Experience

As previously mentioned, the German blessing robot BlessU2—well known in public discourse (Löffler et al. 2019)—was introduced into the interviews as a concrete example for evaluating AI. Responses to it included remarks such as “One could use it [the robot] to distribute hymnbooks, perhaps. But I believe that, when it comes to liturgical use […], it’s not what people are looking for. They come to church in person because they seek human contact” (Interview 15, Position 41). The underlying logic expressed here is that proclamation requires the establishment of a relationship, which is not considered possible in this case. As one interviewee put it, Church actors “are certainly permitted to make use of tools. But these tools are not the ones doing the proclaiming” (Interview 15, Position 53).

In this context, a reference is again made to the android form of BlessU2, which is additionally attributed with AI capabilities. One interviewee remarked “It might help with preparing sermons... but in my view, it cannot replace the preacher, because a sermon must always be connected to the preacher’s personal experiences” (Interview 18, Position 53). A key barrier to the use of AI, then, lies in the belief that it is fundamentally incapable of having or reflecting on experience. At the same time, resistance appears to diminish where personal, positive experiences with AI are present: “No, I don’t think it can replace—but maybe it can complement. There was that prayer robot [...] and the idea was that it could simply pray. I actually quite like that idea [...] but most people reject full replacement” (Interview 21, Position 48). Interviewees often situate themselves within a group like “most people,” signalling how they navigate personal views within broader institutional norms. Personal associations with specific applications are expressed, alongside a resistance to the broader use of AI in the context of ecclesial opinion-forming. This resistance is often linked to concerns about the perceived logic of replacing humans with machines. It becomes clear that many of the interviewees are implicitly referring to what is commonly known as strong AI—a form of technological advancement perceived as posing a threat to human beings and to human culture as it has been known thus far. The notion of substitution is central to this pattern of argumentation.

• Ad H5: Ethical and Theological–Anthropological Misalignment and the Need for a Technology Ethics

It is particularly in the ethical and theological judgements of the interviewees that tensions between established patterns of reasoning and perceptions of AI become apparent. One respondent stated “I really believe we need to clearly define ethical boundaries. […] It’s important to confront pupils and children with moral values, and to make clear what it actually means when we take values seriously—what that means for how we live together” (Interview 12, Position 94). The idea that AI might be programmed with value-oriented parameters or could engage in generative, ethically reflective communication is not considered. Moreover, creativity is explicitly claimed as a uniquely human capacity: “There must be a human nature at the core—made of flesh and blood. A human being can create something, but a machine cannot bring anything into being” (Interview 17, Position 22).

Ethical judgement and decision-making are also viewed as responsibilities that should remain exclusively human: “At least for now, I think it’s important that I, as a human being, make the final decision. And I find it hard to imagine delegating communication entirely to AI” (Interview 15, Position 73). There is a clear call for theological reflection within the field of technology ethics: “We have a highly differentiated political and social ethics... but so far, there is very little competence in tech ethics. That’s beginning to develop... but it’s important to incorporate the theological perspective here as well” (Interview 23, Position 67). There is no need to adopt apocalyptic tones or prophetic social critique: “We need to find an approach that is free from horror scenarios—because we are still far removed from any kind of superintelligence” (Interview 22, Position 14). The interviewees express a need for practical theological and ethical guidelines in engaging with AI. However, they do not indicate that they themselves—or others within the church—should take responsibility for developing them. No specific institutional actor is named.

• Ad H6: Lack of Experience with AI

Church leaders recognise potential in the use of weak AI, particularly in administrative, creative, and organisational domains. They use AI for academic work and are familiar with various applications (Interview 23, Position 12). AI is seen as a tool for improving efficiency and knowledge management: “We’re already experimenting with it […] for corrections, alternative suggestions, and shifts in perspective” (Interview 13, Position 14). AI-based tools are to be used with caution. Work routines that were established before AI became available should be maintained unless a tested and demonstrably better alternative is in place.

5.2. First Indications from the Sentiment Analysis

The sentiment analysis used in this study examines the emotional tone of the previously discussed arguments. While the majority of linguistic expressions in the interviews follow a general and neutral trend, deviations can be analysed by quantifying positive and negative words in the coded arguments. By systematically counting and weighting these predefined terms, an overall score is calculated, which can be either above or below zero, thereby revealing variations in emotional expression. Overall, it is noteworthy that the church leaders interviewed speak predominantly in neutral to positive terms about AI and digital media (see Figure 1). This stands in contrast to the many critical aspects identified through the qualitative content analysis—factors that hinder a proactive integration of AI into church practice or the development of transformative visions for its use. Based on the sentiment analysis, it remains unclear whether this caution and the barriers mentioned are primarily the result of insufficient theological and ethical debate and missing institutional as well as organisational guidelines or whether they reflect limited experience with AI. What can be concluded, however, is that a neutral or even positive semantic framing of AI does not necessarily correspond to an active engagement with it—particularly in terms of individual agency and the willingness to shape its use. It is clear that the thesis presented here very briefly requires further examination.

6. Conclusions and Outlook

The clear majority of Protestant and Catholic leaders accept the use of AI primarily as a tool to support administrative tasks and wish to limit its application to this domain. In this sense, they express an agreement with the use of weak forms of AI.

Although AI is widely regarded in society and culture as a driver of innovation, this dynamic has scarcely reached even digitally engaged experts within the Protestant and Catholic churches in Germany. While a basic technical understanding of datafication processes and AI can be assumed, AI is frequently portrayed and evaluated as a form of intelligence that competes with human agency—thus placing the question of control at the centre of concern.

Overall, it becomes clear that the interviewees rarely speak from a perspective of personal experience with AI in their own professional contexts. The use of AI, its training data, and in particular the processes of modelling and the subsequent moderation by trained experts were hardly addressed in the interviews. There is therefore a lack of organisationally embedded opportunities to explore the potential of AI for ecclesial fields of practice.

At the same time, AI is becoming an increasingly integral part of professional and educational environments. As such, awareness-raising efforts should not be limited to the training and education of future Christian leaders but should also include a systematic reflection and instruction on how to engage with AI. The neutral to positive attitudes towards weak AI expressed by many of the interviewees could provide a starting point for such initiatives.

The interviewees clearly express a call for the regulation of AI use in the church and, ultimately, in society. They are committed to safeguarding the irreducibly human core of Gospel communication, which they do not want to see replaced or shaped by AI. Church communication is described above all as interpersonal and relational work between human beings. Human–machine interactions, as well as communicative artificial intelligence, are rarely discussed.

A theological tradition—traceable within various theological currents—which places the church and theology at a critical distance to (modern) technology appears to have contributed to the barriers identified in relation to AI. Alongside widely shared societal reservations, this tradition may have reinforced hesitations within ecclesial contexts. In response to these challenges, theological reflection has begun to critically reassess the relationship between religion and technology—especially as it has evolved in the West through the Enlightenment and, more dynamically still, in late modernity (cf. Helmus and Riedl 2024; Beck et al. 2022; Harrison 2024).

This study reveals a more cautious attitude towards AI than might have been expected based on the findings of the CONTOC studies, which addressed digitalisation more broadly. At the same time, representative surveys in Germany (cf. Fox et al. 2023) indicate that around 40% of the population feel uneasy or fearful in relation to untrustworthy AI, while only 14% expect positive effects (cf. Fox et al. 2023). The general attitude is characterised by ambivalence—neither clearly rejecting nor fully endorsing AI. There is a high degree of orientation-seeking (cf. von Lindern 2023). The findings on AI and religious leadership can be seamlessly situated within this broader context.

A central finding emerging from the interviews is the lack of agency—that is, a limited capacity or willingness among religious leaders to actively shape the use of AI as part of a wider transformation of the church in a digital age. This is supported by a study of a model digital congregation in Germany, which emphasises that digital transformation strategies must not become ends in themselves but should emerge from congregational concepts that also take account of both regional and supra-regional structures (cf. Sievert and Zeilinger 2024).

It is worth noting that in German-speaking contexts, the concept of “religious leadership” is seldom used and rarely carries a significant weight within church professions. The traditionally esteemed ideal of the priesthood of all believers fosters egalitarian structures in which leadership roles require explicit legitimisation—and in which, moreover, the church’s tendency towards “power-oblivious” modes of action often inhibits the development of such roles (cf. Nord 2021). At the same time, the interviewees unanimously affirm their responsibility for evaluating the use of AI in both ecclesial and wider societal contexts.

Within international practical theology, the term religious leadership is widely used—though usually without reference to digitalisation or artificial intelligence. Instead, leadership is often characterised—without explicitly opposing technological perspectives—as embodied, non-technologised, and oriented towards resistance to societal powers (cf. Cruz 2022).

Many of the religious leaders interviewed appear to be waiting for innovative solutions in church communication to be developed by other actors or to arise from external circumstances. This lack of initiative may stem from the uncertainty regarding theologically legitimate ways of engaging with AI. However, it may also be attributed to a lack of a mandate to develop church-related communication strategies that involve AI. Establishing such a mandate would require institutional and organisational processes that, on the one hand, provide theological and ethical guidelines for concrete ecclesial projects and, on the other, create communicative spaces for building consensus.

Churches are thus challenged to further develop their institutional core through the interplay between Christian tradition and communicative, religion-related innovation. As noted at the outset, this task can draw on international interdisciplinary debates on religion and media—discussions that bring together theology, cultural studies, media studies, and communication studies (cf. Cheong and Campbell 2024; Kurlberg et al. 2025).

The determination of the relationship between human and non-human agency, the capacity to act, and the status of actors will become increasingly important. This includes, at present, questions concerning the conceptualisation of forms of collective, shared, or relational interaction (cf. Schlosser 2019). A more precise understanding of the relationship between human and non-human capacities for action and interaction could help reduce barriers to engaging with AI in an increasingly datafied world. A Christian theological orientation—one that extends beyond the individual—can offer guidance not only to ecclesial actors but also contribute to a value-based ethical debate within the wider society of which the churches are a part.