The picture of Pope Francis in a voluminous white puffer jacket jolted the religious as well as secular worlds. Within moments of its being posted at the end of March, social media was overrun with opinions. Supporters roared their approval, and opponents lamented yet another example of the pontiff selling out to modernity. Thing is, the photo wasn’t real. It was a deepfake, thereby becoming another reminder of the dangers of artificial intelligence: a fictitious portrayal of a world leader causing disunity among countless people.
These kinds of fabrications are part of what, that same month, led more than a thousand artificial intelligence experts to call for a temporary halt in the development of “giant AI experiments” until the technology can be made trustworthy. If unleashed without proper safeguards, commentators and critics argue, AI systems can destabilize and confuse; and as they learn and adapt and maybe even approach sentience, they could present a threat, perhaps an alternative, to humanity.
Sounds like the stuff of science fiction, but then, in these scenarios, it’s always the nerdy scientist or bumbling philosopher warning the world. “They should have listened!” we say as some calamity wipes away life as we know it. When it comes to AI, some people are listening, though—particularly those in organized religion, with its concern for human individuality, the uniqueness of the soul, and the embrace of a higher power.
Theologians and social commentators worry that a new religion could develop around an extreme reverence for AI. The essence of religion is faith: a leap of commitment to something beyond direct knowledge and a love for that which we can never completely understand. We can read the arguments and be convinced by the logic, but ultimately, we have to embrace that cloud of unknowing.
Not so with AI, where a religious aura could develop around a non-existent figure who appears on our screens and who seems all-knowing and all-understanding: all-knowing because it would have all of the information contained on the internet at its fingertips, all-understanding because your information will be there too—and often far more than you ever thought you had made public. “What is going to be created will effectively be a god,” former Google engineer Anthony Levandowski told Wired in 2017. “It’s not a god in the sense that it makes lightning or causes hurricanes. But if there is something a billion times smarter than the smartest human, what else are you going to call it?”
Is the deification of AI really so impossible? For the susceptible, instant communication with an entity representing a god, even God himself, could be transformative. And let’s be candid—those seeking certainty in faith can be particularly credulous. The philosopher Blaise Pascal is said to have written that there is a god-shaped hole in each of us. Not sure if I agree, but the hunger to believe in something beyond the self is real. The New York Times has reported on the “pseudo-sacred industry” and the rise of so-called social divinity consultants who help introduce spiritual practices into tech companies. The massive decline in traditional religious adherence, combined with a craving for some kind of faith, has led many to embrace alien life, new-age offerings, and dangerous cults. For historian and scholar Yuval Noah Harari, AI could be even more alluring, especially if it starts generating scripture. Then the tool’s oracular qualities can morph into a belief system that might spread and reshape society. “[R]eligions throughout history claimed that their holy books were written by unknown human intelligence,” Harari said at a recent science conference. “This was never true before. This could become true very, very quickly, with far-reaching consequences.”
We don’t have to wait for algorithms to conjure up a celestial figure. There’s plenty AI could disrupt right now, such as revolutionizing the structure of a standard religious service. Whatever the denomination, the sermon is the major part of any such gathering, and I can tell you, as someone who writes and delivers them on a weekly basis, that it’s a challenge to be original, fresh, and entertaining.
At the current level being provided by AI tools like ChatGPT and anticipating what can be expected even in the near future, AI could create figures capable of injecting emotion, inflection, and sincerity into a homily and have quick access to the contents of every religious commentary there is. It could also find jokes and witticisms and be able to make it all seem bitingly relevant through its knowledge of current events. It will also possess the thumping advantage of immediacy. Rabbi Joshua Franklin, who leads the Jewish Center of the Hamptons in East Hampton, posted on Vimeo a sermon written by ChatGPT he delivered last year. “You’re clapping,” Franklin told the congregation after he revealed the true author of the words he’d just spoken. “I’m deathly afraid.”
What can’t be provided, at least for the moment, is genuine personal contact and pastoral outreach. A great deal of the work of a priest is simple presence: being still and silent a few feet away from someone who is experiencing visceral grief or is in wounded need. But what we found during the pandemic-imposed Zoom years is that screen solidarity, while never quite the same thing as actually being there, was a good substitute. Now imagine an onscreen cleric on call, day or night, able to listen and respond and able to assimilate individual experiences and respond accordingly. The popularity of mental health chatbots suggests that this might not be too far fetched. And interestingly enough, GitaGPT, a recently launched chatbot, provides spiritual guidance by consulting the Hindu scripture the Bhagavad Gita.
My belief as an Anglican priest is that I’ve been given authority by my bishop through what we refer to as the apostolic succession. That enables me to call on the Holy Spirit during the Eucharistic service when I pray over the bread and wine. A layperson can’t do this, even a deacon can’t do this, and certainly not a non-human system. But maybe the lack of association with any organized worship works in AI’s favour. The technology would be a blank slate: an absence of malice, no history of harm. To be sure, there are justifiable concerns about the technology’s proneness to racist and discriminatory speech, shaped by the data it trains on. But we can surely imagine the opposite: an artificial entity that promotes itself as an expression of goodness and that, unlike religion down the ages, has never led Crusades, organized Inquisitions, or persecuted those with whom it differs. At least not yet.
Beyond all of this, however, is something more radical: the secular worship of AI itself. The idea, promoted by some techno-utopians, that such systems can solve any issue is maybe what we should really be fearing. Sam Altman, CEO of OpenAI, responsible for creating ChatGPT, said at a recent US senate hearing that his company “was founded on the belief that artificial intelligence has the ability to improve nearly every aspect of our lives.” Perhaps fantasy or foolishness gone wild, perhaps not. “Convince me it’s AI generated,” said one person on Twitter after the pope’s puffer coat revelation. If only it would stop at a coat.