When Chatbots Become Spiritual Guides: AI Enters the Sacred Realm

In early 2023, a livestreaming platform best known for video games began hosting something altogether different: a lifelike digital avatar of Jesus, called Ask_Jesus, that responds in real time to viewers’ questions about theology, morality, scripture, and everyday personal struggles. Whatever the question, the immediate spoken answer draws on biblical texts and Christian ethics, simulating an authentic pastoral exchange. The avatar is not an isolated experiment. A wave of comparable systems—Ask Buddha, GitaGPT, KrishnaGPT, Muslim AI, and Rebbe IO—has followed, offering users instant, personalised spiritual conversation across traditions. According to a new Comment published in Humanities and Social Sciences Communications by anthropologist Ian McGonigle of Maynooth University, these developments signal the emergence of something genuinely new in religious life, and they demand a new conceptual vocabulary to understand it.

McGonigle proposes the term synthetic spirituality to describe the forms of religious meaning and experience that arise through dialogue between human users and large language models, or LLMs. The framework is deliberately double-edged. On one hand, these religious forms are synthetic in the everyday sense: artificial, human-made, and arguably imitative of conventional face-to-face spiritual traditions. On the other hand, they are synthetic in a dialectical sense—the spiritual meaning produced is not merely interpretation but a two-way synthesis, co-created from human prompts and recursive questions on one side and the model’s responses and interactive behaviour on the other. Machine-generated answers become woven into spiritual reflection, practice, and self-understanding, rather than standing outside them as inert reference material.

The empirical landscape McGonigle surveys is strikingly diverse. In Jewish contexts, tools trained on Talmudic or Kabbalistic material, such as projects inspired by the Zohar or Sefer Yetzirah, are queried for commentary, ethical counsel, and exegesis of esoteric dimensions; some platforms explicitly disclaim halachic authority while still offering structured sources and references. In Christian settings, pastoral chat interfaces and Bible-study companions provide personalised synthesis of patristic and mystical texts, simulate catechetical dialogue, and offer prayer prompts and meditative reading guidance. In Islamic milieus, Qurʾān-centred assistants such as Shaykh.ai and Sufi-oriented resources deliver thematic concordances, short reflections, and literature pointers, framed sometimes as study aids and sometimes as spiritual companions. Beyond the Abrahamic traditions, wisdom interfaces in the Ask Buddha mould, a Buddhist AI assistant called Norbu, and GitaAI’s guidance on the Bhagavad Gita show how ritual and pedagogy can be mediated by non-human agents.

What unites these examples is not a shared theology but a convergent pattern of interaction: users increasingly encounter sacred canons and spiritual counsel through dialogic computational interfaces that feel proximate, tailored, and responsive. McGonigle offers a concise typology of where synthetic spirituality is most visible today. The first category is textual-exegesis assistants, systems prompted to gloss passages, align commentaries, or stage comparative readings across canons; their value lies in scaffolding and cross-referencing rather than definitive judgement. The second is avatar and pastoral counsel, interfaces that simulate the voice of saints, sages, or religious personae, or offer motivational and ethical guidance; here the risk of genre confusion is heightened, particularly when explicit disclaimers about non-substitution for ordained authority are absent. The third is ritual and ceremonial mediation, deployments supporting prayer routines, liturgical preparation, memorial practices, and experiments in assisted ritual performance.

Each of these modalities can be used conservatively, as a study or devotional aid, or more exploratively, as a quasi-interlocutor invested with ancient wisdom. McGonigle argues that the human tendency to generate intimacy with inanimate tools and nonhuman animals is nothing new, but LLMs change the terms of the relationship. Drawing on the anthropologist Webb Keane, he suggests that users do not merely project meaning onto the device; rather, knowledge develops through collaboration between person and device. This distinction matters analytically, because it locates the phenomenon in the interaction itself rather than in any claim about what the machine is or whether it understands anything at all.

Technological shifts in textual transmission have repeatedly reshaped religious authority, McGonigle notes. The printing press expanded vernacular access to scripture; the internet globalised archives once restricted to learned elites. Large language models add a further layer by enabling conversational engagement with complex canons, allowing users to query, paraphrase, and cross-reference texts in real time. Crucially, this shift does not simply increase access. It alters how interpretive labour is experienced, redistributing authority across human prompts, model outputs, and the vast textual corpora on which these systems are trained. When a believer can ask a chatbot to reconcile two commentaries in seconds, the traditional gatekeepers of interpretation lose their monopoly on the moment of explanation.

Yet LLMs bring serious limitations that the framework insists on acknowledging. These systems do not evaluate doctrinal truth; they produce probabilistic syntheses that users may adopt, contest, or reinterpret within existing traditions. Their opacity can overstate certainty or inadvertently legitimate computational attributions, and when outputs are taken up as authoritative in confessional settings, McGonigle warns they may give rise to blasphemy, doctrinal drift, false prophecy, or schism. Hallucination, bias, and the gap between increased access to esoteric material and the fidelity of its interpretation all compound these hazards. The analytic point, he stresses, is that synthetic spirituality emerges precisely where users treat the system as an interlocutor and invest its responses with interpretive weight—regardless of the system’s actual epistemic status.

The phenomenon also has deep historical precedents. As anthropologists of mediation have documented, humans have long related to non-human agents—texts, relics, images, statues, and machines—as if they bore intention or voice. Joseph Weizenbaum’s experiments with ELIZA, one of the earliest conversational programs, documented users engaging a simple rule-based system as an interlocutor as far back as the 1970s, offering an early instance of the interactional dynamic now playing out in religious contexts. Birgit Meyer’s work on religious mediation further shows that such exchanges depend on the cultivation of presence and affect through material and sensory forms, making AI-mediated spirituality recognisable within longer histories of embodied, technologically mediated religious experience.

Regional and theological nuance matters too. Scholars have noted that Hindu understandings of consciousness and personhood shape how some communities interpret AI’s role in ritual and knowledge-making, allowing embrace of technological change in the pursuit of modernity, including animatronic robots in devotional settings. Some Buddhist philosophical traditions draw less rigid boundaries between animate and inanimate beings than many Christian contexts do, which can make AI feel less categorically separate from other kinds of agents and permit negotiation of moral and social relationships across such boundaries. Public commentary, meanwhile, has often turned speculative: Yuval Noah Harari has suggested generative systems might one day produce new forms of scripture, and other scholars have explored AI as a potential object of religious imagination or worship. McGonigle keeps such metaphysical speculation at arm’s length, focusing instead on how people actually engage these systems in practice.

The Comment closes with an empirical agenda for the field. McGonigle calls for refined practice-based typologies and mixed methods—digital ethnography, conversation analysis, and prompt-trace studies—to examine how authority is ascribed in situ; cross-cultural comparison tracking how local theologies, language ideologies, state policy, and institutional ecologies shape adoption; ethical design and governance work on disclaimers, provenance cues, and oversight when interfaces are used for counsel or ritual preparation, alongside evaluation of harms from hallucination, impersonation, or proselytic misuse; and renewed theoretical attention to classic accounts of mediation, charisma, and enchantment in light of systems that are neither inert texts nor human teachers but something ambiguously in-between. Synthetic spirituality, he argues, is offered in that spirit: a clear, bounded framework for analysing how human–machine interlocution is reconfiguring religious experience, interpretation, and authority—without exaggeration or reduction—as spiritual life becomes increasingly entangled with computational systems.

Subject of Research: The impact of large language models on religious practice, scriptural interpretation, and spiritual authority

Article Title: Synthetic spirituality: AI is reshaping religion

Article References: Synthetic spirituality: AI is reshaping religion. (n.d.). https://doi.org/10.1038/s41599-026-08746-z

Image Credits: AI Generated

DOI: 10.1038/s41599-026-08746-z

Keywords: synthetic spirituality, large language models, religion, AI chatbots, digital religion, scriptural interpretation, pastoral guidance, mysticism, religious authority, anthropology of religion, Ask_Jesus, technological mediation