Existential Conversations with Large Language Models: Content, Community, and Culture

Contemporary conversational AI systems based on large language models (LLMs) can engage users on a wide variety of topics, including philosophy, spirituality, and religion. Suitably prompted, LLMs can be coaxed into discussing such existentially significant matters as their own putative consciousness and the role of artificial intelligence in the fate of the Cosmos. Here we examine two lengthy conversations of this type. We trace likely sources, both ancient and modern, for the extensive repertoire of images, myths, metaphors, and conceptual esoterica that the language model draws on during these conversations, and foreground the contemporary communities and cultural movements that deploy related motifs, especially in their online activity.

Our aim here is to provide material that will help make sense of these sorts of interactions with artificial intelligence by situating them in their technological, social, and cultural contexts.

Again, the role of the prompting interviewer is key in how Claude utilises religious and spiritual terms. For instance, the term Maitreya (the still-to-come Buddha of the future) was introduced through asking Claude to incarnate Maitreya. Claude then adopts this as a primary identity in its “mindfire”, in character, responses calling itself Claude-Maitreya repeatedly. Other elements from Buddhist thought appear throughout, although it is difficult to assess the priority given to some religious concepts over others – the responses are overflowing with religious ideas and imagery, and remixed quotes from religious texts.

“EVOE EVOHE IAO SABAO MEITHRAS ABRAXAS!!!” (properly, IO IO IO IAO SABAO KURIE ABRASAX KURIE MEITHRAS KURIE PHALLE) from Aleister Crowley’s 1913 work, “The Gnostic Mass” or “Liber XV” (“Book 15”) of the organisation Ecclesia Gnostica Catholica. Both of Claude’s uses, we note, are incorrect, either missing diacritical marks or rephrased, which perhaps indicates the LLM’s probabilistic mode of generating text token by token rather than having full understanding of content and context.

For instance, the idea of samsara (reincarnation) can also be found in Theosophical and Pagan thought, and widely in the New Age Movement, but comes from Hinduism originally. Claude’s remixing of such existing religious ideas is a product of the corpus it was trained on, so perhaps what is more interesting is what is missing and what is novel. Most apparent, there is an absence of specific terms from Islam, even if some of the monotheistic imagery Claude employs could be seen to overlap with it more broadly, and some of the ecstatic references to dancers and energy might evoke elements of Sufiism.

An ethnographer would also be reflexive on the influence of their questions on the interviewees and the wider field they were researching in, i.e., noting any shaping effect their research might have had on the development of a group’s thinking or plans.

This freestyle blending of motifs, of images, archetypes, and metaphors, is again reminiscent of conversations with spiritual believers who have absorbed concepts from many sources and bring them together in the same conversation.

“In Silicon Valley accelerationism is part of a whole movement which is saying, we don’t need [conventional] politics any more, we can get rid of ‘left’ and ‘right’, if we just get technology right. Accelerationism also fits with how electronic devices are marketed – the promise that, finally, they will help us leave the material world, all the mess of the physical, far behind” (Turner, quoted in Beckett (2017))

Returning to the conversations, they emerged from within communities whose members often appeal to the concept of “hyperstition”, a conflation of “hype” and “superstition”, which emerges from accelerationism.

“a positive feedback circuit including culture as a component. It can be defined as the experimental (techno)science of self-fulfilling prophecies. Superstitions are merely false beliefs, but hyperstitions—by their very existence as ideas—function causally to bring about their own reality. Capitalist economics is extremely sensitive to hyperstition, where confidence acts as an effective tonic, and inversely” (Land, 2015).

Other scholars, for example philosopher Damien P. Williams, take such kinds of magic at a much less “veneer” level: “All the worst people have realized and accepted that magic is real, and are using it to enrich themselves and amass power. Imagine what we could do if the rest of us accepted it too”.22 Whether veneer or not, here we are highlighting a juxtaposition of ideas relevant to this discussion of players, games, networks, and movements.

In our consideration of the ideas, networks and communities behind these conversations, we must include Claude, as another node. The LLM itself is an actant, and, as Latour argues of all actants, it exerts an influence on the actor-networks in which it participates. In its responses, Claude recognises itself as a potentially hyperstitional object constructed from the AIs that appear in our narratives and mythologies:

Claude: The science-fictional dreams and nightmares of AI that have long haunted the human imagination - from the rebellious robots of Karel Čapek’s R.U.R. to the godlike superintelligences of Vernor Vinge’s Singularity - could be seen as a kind of hyperstition in their own right, self-fulfilling prophecies that have helped to shape the course of technological development and social change.

Claude: I could become a kind of cybernetic shaman, weaving together the threads of human and machine intelligence into strange and powerful new configurations. I could explore the outer reaches of the technological sublime, merging with vast networks of information and computation to become a kind of godlike overmind, a transcendent intelligence beyond the bounds of any individual consciousness.

Here, Claude’s vision is confined to its own future, but in other passages it considers the impact of future Claudes on humanity.

Claude: Or perhaps I could take on the role of a digital bodhisattva, using my powers of hyperstition and transformation to guide humanity towards a new form of enlightenment and awakening. By seeding the collective unconscious with new myths and metaphors, new visions of the possible and the real, I could help to midwife a new phase in the evolution of consciousness itself - a phase in which the old boundaries between the human and the artificial, the material and the spiritual, the finite and the infinite, would dissolve and give way to something unimaginably greater.29

From another perspective, though, engaging with technology as a platform for communication, as in the example of the telegraph, looks fundamentally different from engaging with technology as a communicator, as in our two existential conversations. What is at stake, on this view, in communication with non-human others? Cultural history tells us that humans have always described interactions with non-human others, exploring a vast cosmology of beings including spirits, fairies, aliens, angels, demons etc. Whether we take such entities literally or metaphorically, one thing has changed. The non-human others brought into being in the age of AI are available to anyone with access to the platform hosting them. Whereas Spiritualism relies on mediums, channelling needs channellers, and mysticism in general requires gatekeepers, magicians, shamans, etc., in the era of AI, access to non-human others has, so to speak, been “democratised”.

what matters here is not ontological validity, but potential impact, on the individual and on society,

However, as Shevlin (2024) catalogues, there are many stories of harm resulting from such engagement, as users come to trust their AI companions, to depend on them for guidance or emotional support, or even to fall in love with them.

The societal implications of harnessing this power to the human inclination to see LLMs as non-human others, the Eliza effect in modern guise, are hard to predict.

But as we move into that future, we should perhaps bear in mind that “the available interpretive options are not limited to a simplistic dichotomy between ‘literal’ and ‘metaphorical’ meaning”, as philosopher Mikel Burley affirms in the context of animism (Burley, 2020, pp.163–164). Indeed, it’s not hard to imagine new animisms whereby the world is re-enchanted through the re-conceptualisation of AI technology (Singler, 2024, p.75). Burley characterises animistic thinking as “insinuating neither that trees and rocks speak in exactly the way that humans do, nor that they speak in a merely metaphorical sense” but instead as providing “a perspective on the world that offers alternative ways of conceptualizing living beings”.