Are AI Gurus a Good Idea?
Original paper by Jacob Copeman. This article was created by AI and is a simplified version of section 3.1 in the original paper.
A rich and growing body of work now examines AI and religion, both conceptually and ethnographically (Kimura 2021; Singler and Watts 2024; Rähme and Prohl 2025). Yet even the fullest recent survey—The Cambridge Companion to Religion and Artificial Intelligence (Singler and Watts 2024)—never discusses the guru at any length. Geraci and Kaplan’s chapter on Hinduism, for example, does not mention the figure at all.6 So far, then, this literature has not asked what is at stake when the guru becomes algorithmic. That gap is an opportunity. AI-mediated guruship is an unusually good site for studying how code reshapes authority and intimacy—part of a longer pattern in which existing forms of guruship keep being remade. The idea echoes Grimaud‘s (2021) call to ‘make your own gods’: his account of new deities built out of robotics kits speaks directly to the speculative devotion visible around AI gurus. Here the phrase becomes ‘make your own gurus’—figures that users not only consult but increasingly help produce and circulate. Rather than simply receiving an AI guru, users can shape how it develops through their prompts and their choices about what to keep and share, effectively helping to train its responses. As Baas (2025) notes, generating output with a tool like ChatGPT reveals what users value, and that information feeds back into how the model is tuned. With AI gurus, this can mean devotees gradually scripting the very authority they hope to receive. Like Grimaud’s deities, the AI guru is neither wholly invented nor simply inherited; it is an unstable artefact of desire and design.
What matters, then, is the set of circumstances under which an AI guru comes to be treated as authoritative, and what it actually does. Questions of authenticity or artificiality become part of that inquiry rather than the place it starts. Such encounters produce their own emotional intensities and their own architectures of trust and risk. What happens to darshan (the act of seeing and being seen by the divine) when the guru is met through a clone, and under what conditions does that meeting get recognised as darshan at all? The prospect of performing seva (devotional service) toward a recommendation algorithm unsettles inherited devotional categories in a similar way. How do users tune themselves to sources of charisma that are neither fully human nor fully machine? Following Grimaud and Wacquez (2023), analysis here should avoid both technophobic alarm and techno-utopian celebration, staying open to the mixed feelings and unresolved tensions in how devotee-users engage.
I call these processes Transhuman Guru Worlding. The term is meant to capture how new realities are generated when a guru’s charismatic legitimacy is built through large language models (LLMs). Building on Demuro and Gurney (2024), who describe LLMs as carriers of ‘transhuman language ontologies’, this approach treats AI guru-bots as world-making participants rather than passive tools.7 Through repeated back-and-forth exchanges with users, they can generate hybrid guru domains—guru cosmoi—that develop their own conventions of address and expectation, and within which guidance is experienced as live and authoritative. These spaces are neither purely human nor purely machinic; they are co-produced by training data and devotional practice. The wager is to treat algorithmic guruship as both imitation and origination: it emerges by copying existing forms while also bringing new devotional worlds into being. Transhuman Guru Worlding names the slow drift into something new that occurs when devotional texts and interface design together produce fields of guidance that feel ‘more than simulated’. These exchanges do not create sealed-off spiritual universes. What they cultivate are distinctive guru cosmoi: relatively stable fields of expectation and address that become recognisable through repeated interaction, and within which users act as if the guidance were live and authoritative (Reed 2021). The term draws attention to how these fields form, how they are policed or pruned, and where their edges fray. It also distinguishes two scales of the same process: worlding names the generative dynamic, and guru cosmoi the relatively stable formations that emerge from it.
One objection is obvious. Calling guru-bots world-making actors may seem to grant too much agency to a next-token predictor—a language model that produces text by repeatedly guessing the next unit of text. So it is worth being precise. ‘World’ here refers to the fields cultivated through sustained interaction: fields within which users act as if guidance were live, and which have describable conventions and boundaries. This is an ethnographic claim about users in sustained interaction with AI systems, not a metaphysical claim about the systems themselves. Keane (2025) would remind us that these emerging cosmoi rest on historically layered signs and social forms that users already inhabit. AI does not generate worlds from nothing; it draws on and reanimates existing repertoires while adding new kinds of responsiveness and elaboration. The strongest ethnographic accounts of Transhuman Guru Worlding would therefore specify the interaction order (Garfinkel 1967; Goffman 1983)—turn-taking, terms of address, repair, the choreography of prompt and response—through which the cosmos is co-produced. Seen this way, simulation—the MetaGuru discussed below—becomes methodologically useful, because it lets the interaction order itself be varied experimentally, making it possible to observe the conditions under which guru cosmoi form or fail to form.
Discover more from Harmonist
Subscribe to get the latest posts sent to your email.
Related Posts
Mystic Poetry and Yogic Enlightenment →
Śrīla Jīva Gosvāmī →
The Love of Sri Radha →
