For many, AI is a religious experience (rerun)

This episode originally aired on Aug. 9.

Artificial intelligence can feel abstract, so we’ve come to depend on certain narratives to try and make sense of it all.

Some of the language we use to describe AI and our interactions with it is rooted in religious ideas. Are you bracing for the apocalypse? Have you been blessed by the algorithm? Have you consulted a Robo Rabbi lately?

The deification of AI, whether it’s done consciously or not, is something Beth Singler studies as a professor of digital religions at the University of Zurich in Switzerland. Marketplace’s Lily Jamali spoke to Singler about religious tropes in the narratives we consume and share about AI.

The following is an edited transcript of their conversation.

Beth Singler: This religious thinking comes from the kind of existential fears that people have about the development of AI, where it’s going and where it might get out of control in particular ways. Those sorts of ideas, those imaginaries, they replicate sometimes religious narratives and tropes. We’ve already seen mainstream media talking about godlike AI and Elon Musk talking about AI summoning the demon. There’s this idea that this thing could get out of control, and the language that’s used pulls on those existing religious ideas. These ideas usually come from monotheistic religions, like Christianity, Islam or Judaism, and the idea of a singular powerful entity that is, in some ways, omniscient, omnipotent and we hope omnibenevolent. But there’s a lot of conversation about how we make AI to be good, to share our values. And that’s where the existential fears come in.

Lily Jamali: And what do you make of that framing? Does that make sense to you?

Singler: I think there is a common conception that AI is, that it is a modern thing and religion is a not-modern thing. Religion is too traditional, too in the past, and it’s something that we should leave behind. Actually, religion and AI are both modern and they both are entangled with each other. You get people thinking of AI religiously and you get people doing religion with AI and they’re interacting with each other in various ways. So, when you get these conversations about AI and how it might become superintelligent, or the thing we call “the singularity,” this future AI that’s going to be all-powerful, you see people drawing on their cultural context and using the terms and the ideas they’re familiar with, even if they aren’t at all religious. There’s that familiarity with these ideas of a godlike being that oversees the world. Other concepts like heaven and the afterlife tie into people’s conversations about digital immortality and AI. We’re basically using our cultural resources to try and understand this thing that’s almost happening to us, more than is something that we’re creating.

Jamali: What have you noticed about the way people think about the algorithms behind social media platforms like Instagram and X, the app previously known as Twitter?

Singler: In this space of uncertainty, you get people talking about how the algorithm has done certain things, made certain decisions and people expressing the feeling that they’ve in some way been blessed by the algorithm. I wrote a piece of research specifically on people on Twitter using this expression. Where it came up most commonly was around spaces where there are automated decision-making systems like the algorithm behind YouTube that decides what you’re going to see next. Or if you’re working in a gig economy job as a driver for Uber or Lyft, you might feel that if you’ve had a good day, then the algorithm has blessed you. It’s this idea that the algorithm is making some sort of choice about your future because you’re virtuous and so you’ve been blessed by the algorithm. In reality, all that’s happening behind the scenes is that there’s been some very distinct decisions made by humans about the values put into those algorithms and what sorts of things should be optimized. That’s not always very clear, so we tell these stories to ourselves about being blessed by the algorithm.

Jamali: What are the harms of projecting religious ideas onto AI?

Singler: There is a concern that by seeing these things as godlike when they’re not, it’s obscuring what’s actually happening with the technology. If we put too much trust in AI and we see it as not just an agent, but actually a superagent similar to a god that has the capability to control and guide our lives in particular ways, that will obscure the humans who are actually making the decisions behind the scenes. In theory, there’s nothing specifically harmful about people creating new mythologies around AI, but if that means that we don’t notice the actual changes that are happening in society and the inequalities that are exacerbated by artificial intelligence, then that’s where the harm comes in.

Jamali: So how is AI being incorporated into existing religions? And to what end?

Singler: A lot of the time, religions are very early adopters of technology because they see the affordances and the abilities of new technologies as very useful. The Church of England in the United Kingdom developed Alexa skills that you can upload to your existing Alexa that allow you to ask Alexa to recite the Lord’s Prayer or to say grace before a meal. Now, in the press conference, they were very clear that this isn’t Alexa praying for you, but it’s meant to teach you the words. Another example is a system that a computer scientist named Lior Cole developed called Robo Rabbi. You ask questions of it much like you would ChatGPT, but the content it has been trained on is rabbinic law and elements of the Torah. And there’s also embodied forms of religious AI that provide services. There’s Mindar in Japan, which is a Buddhist robot, and there’s various other theomorphic robots fulfilling religious tasks in different locations.

For some people, this is perhaps seen as a bit of a gimmick, a performative thing to get people interested in conversation about religion and AI. But there are people who think there is a path towards religious experiences through technology. This is a long-standing thing. You can go back to the origins of the internet and some of the early technopagan groups saw the internet as a cyber-pathway towards enlightenment. It’s not, it’s not entirely new in that sense. It’s just that we’re getting more and more used to these new uses of AI.

Jamali: Do you think that AI could bring about new religions and practices?

Singler: Yeah, absolutely. There are various different ways that could happen. AI itself has already inspired some people to purposefully create AI religions that are focused on either the singularity or to try and work towards that. People are also using AI to develop new texts, and they think there might be a way to find new religious inspirations and innovations through generative AI like ChatGPT by asking it to write new religions from scratch. So, there’s a sort of possible space for increased religiosity. But there’s also concerns around not just the stimulation of new religion, but the simulation of religion as well. The historian Yuval Noah Harari has written a lot about the future of humanity and AI. He suggests AI might be employed to create fake religions that inspire people to do particular things because it knows how to control humans. So, there’s that more cynical, pessimistic side as well.

More on this

In our conversation, Beth Singler said the influence of religion on our understanding of AI also works the other way. Advanced technology is already influencing how some people practice religion, like those on the “cyber-pathway towards enlightenment” she mentioned.

Digital theologian Adam Graber wrote in Christianity Today about the rise of large language models that help guide believers through religious texts and the potentially big and very modern implications for how people relate to these ancient scriptures. BibleGPTs — they’re a thing!

The Los Angeles Times tried to tackle the question of whether religion can actually save us from AI in March. The short answer: probably not.

The article brings up the genie in the bottle concept, which is derived from the “djinn” — a figure in Muslim and Jewish folklore. The genie can grant you a wish, but with one catch: Once it’s out, you can’t put the genie back in the bottle.

Perhaps the djinn, which originated in these spiritual traditions, serves as a cautionary tale worth noting as we ponder the ethics of AI.