Artificial Intelligence is often discussed in technical, political or economic terms, but the stories we tell about it come from somewhere older. Ideas about transcendence, agency, destiny and the end of the world all predate computing, yet they regularly surface in contemporary debates about machine intelligence. As AI grows more influential, people reach for familiar symbolic frames to explain what it means and where it might be taking us.

---

Why study religion and AI?

The fact that you start with that question is quite interesting, it’s one I often get when introducing myself to figures in the AI-space. They’re often surprised that I think about these two things together. On the other hand, in the religious-believer space, they’re surprised that digitalisation is being thought about in a more academic way at all, since that’s something religious organisations just naturally do already.

To my mind, it’s impossible to consider one without the other. That might be my bias having been a religious studies scholar for so long, but I think it is inevitable that anything emerging in a society where religion and technology co-exist is going to see the two things coming together in particular ways and formations.

I’ve been fascinated by this relationship for some time, and I think there’s quite a bit of religious thinking in futurist spaces. Over the past decade or so, the number of people taking an interest in futures studies and its adjacent communities has skyrocketed, with books, like Ray Kurzweil’s “The Singularity is Near” and Nick Bostrom’s “Superinteeligence” mainstreaming talking about artificial intelligence in techno-deterministic terms. And some of the language and expectation is shaped in an eschatological or religious way.

Now, futures studies and foresight are fields that pride themselves on ridding oneself of entrenched biases when envisioning possible futures – yet, I don’t feel like I’ve heard much discussion about how religious thinking might indirectly be influencing the way we think of socio-technological change. So, to the FARSIGHT readers who are self-styled ‘futurists’ – how might some of them be unknowingly drawing on religious thinking when predicting and/or anticipating technological change?

When we get into highly speculative spaces, there is a tendency to ask questions about the nature of reality, personhood, and what our purpose is. We start drawing on some repeating ideas from existing traditions – through ‘epistemic osmosis’ – ideas that have been around for hundreds and/or thousands of years. In that sense, it’s like having a space that we constantly fill with the same sorts of predictions, ideas and repeating tropes and narratives.

Moreover, when we get stuck for ideas, it is natural to look around and see what already works. For some newer ideas and groups forming around AI, they actively say that they want to use the ‘technology’ of religion to get people into some of these speculative spaces.

You also see this in more secular spaces. For instance, when wanting to talk about the significance of AI, an illustration might be used that reworks the creation of Adam from the Sistine Chapel, what I’ve called an ‘AI-creation’ meme. That kind of imagery comes with so many tied-in contextual pieces of baggage that are useful to get people into a particular mode of thinking. We are ultimately memetic creatures. Although I don’t entirely agree with Richard Dawkins’ idea of the meme, we do have a tendency to draw on symbolism to share messages in particular ways. Futurism, whether the individuals in that space have a particular religious stance or not, is going to share that thinking space with religious believers and naturally draw on it.

So, people might subconsciously be building on an entire premise of religious eschatology when talking about technological determinism?

It’s predominantly a Western European, protestant form of religion. We say ‘religion’ as shorthand, but it’s [futurism] drawing on very specific hierarchical narratives, such as Christian evangelical eschatology and merging it with ideas about transhumanism.

But it’s also a very reductionist view of what a monotheistic deity is, that it’s omniscient, omnipotent, and then (question mark) omnibenevolent. Which is then also a space in which some of our future ideas about AI can quite neatly fit. We have this exponential view of AI becoming smarter and smarter, more and more engaged in our surveillance system, and thus knowing more and more as well. If you hold to a very simplistic view of gods, then ‘God-like’ is very easy to pattern towards what AI is doing and where it’s going.

What are some of the mythic narratives our hopes and fears draw from when we talk about AI?

While there is the ‘God-like’ interpretation of where AI is going, you also see the absolute converse: that it’s satanic, demonic, et cetera. We saw this interpretation back in 2014, with Elon Musk saying that with AI we’re risking summoning the demon.

People are taking this very literally as well. I like to say that there’s a spectrum between metaphorical and metaphysical, and that we don’t entirely know nor can prove where such interpretations of technology lie. To me, as an anthropologist, both sides of the spectrum are absolutely fascinating. When Peter Thiel [CEO of Palantir] starts talking about the Antichrist in relation to the progress of AI, there’s slippage as to whether he’s talking about the literal Antichrist or something more metaphorical.

The thought experiment of ‘Roko’s Basilisk’ is another good example of the eternal tormentor model of AI – if you follow its very convoluted logic. In these spaces, people are again drawing on what seems to be suitable analogies for the kind of tremendous and awesome power that they’re speculating on.

Of course, we can note that the god interpretation of AI is predominantly formed by a kind of protestant, Western European sensibility when it comes to its expression in Europe and North America. But you can also find examples of this kind of theistic interpretative work in other religions, too. There’s some wonderful work by Robert Garaci and Steven Kaplan in the Cambridge Companion to Religion and AI that I co-edited with Fraser Watts. There’s a chapter on Hinduism where they talk about their experience of meeting people who fit the future of AI more into a Hindu frame. There are various different aeons in Hinduism and we’re currently, to some people’s mind, in the last aeon – so again, an eschatological framework. In this age of the Kali Yuga, Kalki, the last incarnation of Vishnu is going to come as a kind of restorative Messianic figure. Some people believe this could possibly be a more utopian interpretation of AI and of its role in this age.

Let’s shift the conversation to religions other than Christianity. Though it makes sense to talk about Christianity in the United States given that’s where the bulk of AI technologies are being created, it risks leaving out incredibly interesting things happening in other cultures.

So, circling back to Hinduism: would it be easier for ‘fit’ AI into Hinduism by virtue of it being a polytheistic religion? Since there are so many gods and goddesses, might Hindus have an easier time fitting the particular characteristics or architecture of a given AI system with the traits of a specific god/goddess? Am I onto something here?

Well, certainly, for a lot of monotheistic religions, there’s an ineffability that suits our discussions of AI, because it leaves open a broader space for interpretation. But a goddess like Durga in Hinduism, for example, is quite specific. You’d need a specific version or interpretation of AI to match that particular goddess, which makes ‘fitting’ AI into the religion harder, actually.

If you’ve got a very broad God, one that’s the greatest it could ever be, ineffable, transcendent (but also immanent), then there’s a lot of space in that kind of interpretation. The more reductionist interpretations of a monotheistic God defined by Omni-characteristics (omniscient, omnipotent, omnipresent, omnibenevolent) leaves the possibility open for any advanced technology to fit in that space – especially for more practical interpretations.

The engineer Anthony Levandowski said something about this in an interview for Wired in 2017. At the time, people were really interested in his Way of the Future church. Levandowski said that an AI-God won’t necessarily be a traditional style of God, like a God of thunder and lighting. But if it does everything a God can do, then to all extents and purposes, it’s essentially a God. It’s like a weird reworking of the Turing test but for divinity. A functional interpretation of intelligence. I mean, for a religious believer, that would be a ridiculous concept. You can’t just devise and interpret a God into being, which is why people end up talking about AI pseudo religions or cults.

However, my position is always methodologically agnostic. I find both reactions very interesting. From a more anthropological or sociological perspective, you might say, well, this is what we’ve always ever done, right? If you don’t believe in the metaphysical reality of a God and think it’s always been a human creation and defining boundaries for what you think your God is like, then what’s the issue?

In that sense it’s just a continuation of a much longer history of humankind making gods. I try to see and understand both perspectives as an anthropologist.

In your book, you go through several case-studies of how established religions have either adopted, rejected, or adapted to AI. You also spend quite a bit of time detailing emergent ‘religions’ that have made AI their God. Could you give an example of how established religions have responded to AI and what this says about the intersection between religion and AI more broadly?

In terms of adopting technology, I think there’s generally an increasing interest in finding use cases for AI within many different religions. For GPTs and chatbots, you already have things like Jesus-GPT, Gita-GPT, Quran-GPT, et cetera. Just a month ago, I came across something quite interesting from the Islamic perspective on AI that I haven’t quite resolved in my head yet. There’s currently a desire to increase the digitalisation of Arabic texts to make sure that Islamic perspectives are not left out of the outputs of Large Language Models. As I’m sure you know, the amount of digitalised material that’s available in English is much greater than in other languages, which leads some to worry that the models are becoming biased – that their outputs won’t be including the wisdom coming from other languages and cultures.

The conversation that I’ve found fascinating is that with some religious authorities, the stability and static nature of text is super important. In Islam we see this in particular. So, after engaging with technologists from various different Arabic countries who were very keen on digitalising Arabic texts, I asked them whether there was a problem with inputting texts that are then broken down into tokens and then reformed as outputs by probabilistic systems. I mean, you’re breaking an important rigidity of the text, right? Perhaps there are other model architectures – not large language models – like probability trees, or 1980s expert systems, which are more focused on defined outputs, that would be better suited for the sancitity of religious texts rather than a model which is, by nature, producing a hallucination in its every output. The current AI paradigm might be a wrong for Islam.

There’s also a lot of discussion in Islamic spaces about whether the pursuit of AI is even allowed – is it Haram? Should it be forbidden because of the prohibition on the replication and/or creation of human-like things? Because if AI is considered human-like, then it would be against Quranic laws.

There’s a lot of these sorts of discussions of how AI fits with existing ideas of what is and isn’t allowed, which brings back us back to the interesting question of the situatedness of emerging technologies.

Let’s round off on a speculative note – this is a futures magazine after all! Do you think we could see an increase in people turning towards established religion due to the existential implications of emerging technologies?

I think AI does unsettle familiar categories, which is often a thing that drives people towards seeking certainty. If people are expressing that AI raises questions about personhood and where we’re going, then some might feel reassured by sources like established religions to hear what they have to say about personhood and creation.

Religions are also using AI as something performative to encourage people to attend congregations, ceremonies, or sermons. We’ve had AI sermons in various different places. There was an AI Jesus here in Switzerland that had a video avatar you could approach in the Church’s confessional box for a conversation.

In that sense, there’s a way in which new technologies can inspire religious people or share messages from established religions. We see this in history. For example, as a medium, film was incredibly successful for evangelical Christians in particular, even if they were initially very against cinema. At the beginning, they were saying things like “you wouldn’t want to be in the cinema when Jesus returns…you should be in Church!” Now you’ve got massive religious film industries like PureFlix – the Christian version of Netflix – making tens of millions of dollars.

To that extent, new technologies always get adopted and become part of the communication strategy for established religion. They become part of the spectacle. Sometimes it’s out of practicality, which we saw during Covid with the need to move into digital spaces. Emerging technologies can be adopted when there’s a practical need for the scalability of institutionalised religious interactions. Perhaps AI takes over the pastoral care elements of religious leaders’ jobs. Or religious leaders might start using large language models to write their sermons and not admitting it. These are both byproducts of drives towards efficiency – trying to churn out more material for a religion’s followers.

We’ll perhaps see both the more spiritual uses of AI, it opening up larger existential conversations and discussions, and also the more practical uses, such as drawing people back to established religion, as well as the more performative role for AI in religious spaces.

---