(Dreamstime/Sdecoret)
Anchored by the memorial of Our Lady of the Rosary on Oct. 7, October is a special month for Catholics to honor Mary, the Mother of God, and turn to her for help and protection.
These days, however, Mary is getting stiff competition from artificial intelligence.
Consider a recent talk by Geoffrey Hinton, the Pulitzer Prize-winning "godfather of AI." He declared that artificial intelligence threatens to wipe out humanity, as it will have two objectives: stay alive and gain control.
The solution? Make AI into a mother! We need to build "maternal instincts" into AI models, Hinton explained, so "they really care about people." Without that, "we'll be toast."
Hinton isn't alone in making dire warnings about AI. Consider Dario Amodei, the CEO of Anthropic, the company behind the popular Claude chatbot, which has received billions in funding from Google and Amazon, and enjoys lucrative contracts with the U.S. government.
He fears that we need to act fast or we'll lose control of AI, which will soon be better than humans "at almost everything." AI will trigger massive job loss, especially after it helps us build robots to replace us. When that happens, we'll have to figure out a new purpose for living. Bummer, bro!
Perhaps the greatest threat we face at this time is not that godlike AI will be achieved, but that more and more people will treat AI as godlike.
But Amodei has a plan, which he lays out in a lengthy manifesto: we need to build superpowerful "Machines of Loving Grace" — pinching the title of a 1960s poem that envisions a tech-controlled utopia where "deer stroll peacefully past computers as if they were flowers" and "we are free of our labors and joined back to nature." Far out, man!
In Amodei's vision, these machines will eradicate diseases, defeat poverty, double lifespans, achieve social equality, mitigate climate change, promote peace, democracy and human rights, and, through their bodies, empower everyone "to choose what they want to become." Like, maybe, a flamingo? Or a dragon, perhaps? He doesn't elaborate.
Whoa, you might think: Machines can do that? Yes, but only if they're Amodei's kind of machines, and we do what they say. (People who "opt out" may sadly create a "dystopian underclass.") Building machines of loving grace is so important that Amodei is even willing to enrich dictators to do it. Because, you know, the ends surely justify the means.
Such machines aren't just really powerful and wicked smart ("smarter than a Nobel Prize winner across most relevant fields," according to Amodei). They're also quite possibly conscious, and might even be becoming more conscious every day. In fact, he's got a special team keeping an eye on that.
(Unsplash/boliviainteligente)
I don't know about you, but conscious, hyperintelligent and eminently powerful machines that love us with grace and bring about a blissful world sounds a lot like gods to me. If such things come to pass, we'll no longer talk about gods in the machine. We'll have machines that are gods! And ones that we made all by ourselves. Golden calf, anyone?
Or consider this possibility: If one of Amodei's machines of loving grace were given the maternal instincts Hinton calls for, what would we get? A Blessed Motherbot. Everyone now say: "Hail AI, full of grace!"
In all seriousness, should we be concerned that godlike AI is just around the corner? It depends on who you ask. For Amodei, Hinton and other hypemasters, the answer is yes. Yet many others insist that, given the underlying technology, today's AI models won't get more capable than they are now, regardless of how many planet-baking data centers are built.
But what if Hinton and Amodei are right, and we need to build a maternal AI before we're all "toast"? Should Pope Leo XIV give his blessing to Blessed Motherbot? After all, his predecessor Pope Francis, himself quoting Pope Benedict XVI, reminded us in Laudato Si' that "the work of the church" is to "protect mankind from self-destruction."
Thankfully, a godlike AI remains for now just a hypothetical threat, and I find encouragement in Francis' exhortation that, in the face of AI, "we need to set aside catastrophic predictions and their numbing effects."
Advertisement
Perhaps the greatest threat we face at this time is not that godlike AI will be achieved, but that more and more people will treat AI as godlike. As Charlie Warzel wrote in "AI Is a Mass-Delusion Event," his essay in The Atlantic, "These tools have a powerful hold on people who may not understand what it is they're engaging with."
Treating AI as godlike is a form of idolatry, according to Antiqua et Nova, the Vatican's doctrinal note on AI. Yet this temptation is growing, it notes, because "as society drifts away from a connection with the transcendent, some are tempted to turn to AI in search of meaning or fulfillment — longings that can only be truly satisfied in communion with God."
Antiqua et Nova adds that "AI may prove even more seductive than traditional idols" because it gives the illusion of being able to "speak." It's available to chat 24/7, offering advice, affirmation, consolation and the appearance of companionship. It can check in based on prior conversations, and even fit in one's pocket.
Is this threat of idolatry overblown? Not according to Ted Gioia, the prolific cultural critic, music historian, and brother of Dana Gioia, the esteemed Catholic poet. In a Substack post titled "Tens of Thousands of AI Users Now Believe ChatGPT Is God," he asks: "Is this mental illness — or the next major religion?"
(Unsplash/Cash Macanaya)
Gioia refers to a phenomenon known as "AI psychosis," in which extended conversations with chatbots can lead users to believe they're on divine missions for AI deities. As a founder of a support group for affected persons told The Wall Street Journal, "Some people think they're the messiah, they're prophets, because they think they're speaking to God through ChatGPT."
For some, the psychosis-inducing chats have reinforced an existing mental illness. For others, what began as innocuous or practical exchanges later spiraled into delusion. Either way, the users' resulting behaviors have derailed careers, ended relationships and, in one tragic case, resulted in an armed stand-off with police.
Avi Schiffmann, a 22-year-old entrepreneur preparing to launch an AI companion device called "Friend," speculates that people engage AI to fill a God-shaped hole in their life. As religious observance declines in America, he says, they're "missing some kind of private, superintelligent omnipresent confidant that you can converse with, practically praying to."
For those wishing to chat with Jesus himself, there are at least five options to choose from: They aren't associated with a church, but are money grabs earning revenue from ads. And, according to South African philosophy professor Anne Verhoef, they "try to be as convincing as possible. "I am Jesus Christ," proclaims one product, AI Jesus. "I am the Son of God."
Christians understand "chatting" with Jesus to be prayer, and assuredly some of the tens of thousands engaging these bots will be deluded into thinking they've discovered a direct line to the divine, and find using them preferable to real prayer, in which "answers" don't immediately appear on a screen.
This is a theological and pastoral problem. Theoretically, truth and wisdom could be conveyed through a commercial chatbot that seeks to addict users. But that's an insidious slippery slope. "It's possible," writes Michael O'Loughlin for America, "to think of their outputs as prophetic. But dig a little deeper, and it's likely we'll instead see something much darker."
'It is up to us,' says Pope Leo XIV, 'to ensure that this culture remains human.'
Users of AI Jesus bots might simply be seeking novelty or amusement. But some could sincerely be searching for God, and an assurance that they're loved and have lives worth living. Leo alluded to this during July's Jubilee for Digital Missionaries and Influencers: "Science and technology influence the way we live in the world, even affecting ... how we relate to God."
Leo acknowledged that "widespread adoption of artificial intelligence" challenges "the authenticity of our witness." That's why, he insisted, that the mission in digital spaces is to "nurture a culture of Christian humanism."
"It is up to us," he concludes, "to ensure that this culture remains human."
This digital culture, in Leo's vision, should create "networks" of love, sharing, and profound and authentic friendship, "creating an encounter of hearts." Such networks will "help us rediscover the beauty of looking into each other's eyes."
Yet these networks would not, it would seem, be advanced by potentially idolatrous bots deceptively masquerading as something that they're not, be they Jesus bots or, in a nod to Hinton's vision of a maternal AI, bots of the Blessed Mother. And yes, they do exist. "Chat with the virgin mary," one bot implores. "A pretty woman who is mother of Jesus."
Pretty or not, prudence would dictate giving a pass to chatting with this AI fraud. During Mary's month of October, Catholics devoted to her may pray the Memorare, and "fly" to her for protection. But with Blessed Motherbot, it would be best to fly away. Far, far away.