The other day I went to a really interesting talk by Carl Öhman (Uppsala University) about applying religious critique to AI.
Below are my notes from the presentation. As usual, these were taken quickly during the talk and I may have misheard or misunderstood parts of this. I’m mostly putting these here for my own benefit and later review.
AI as a god in the anthropological sense
Religious life originally emerged from ancestral authority. In small societies, people look up to their ancestors for guidance. But over time, as ancestors become too numerous and too distant, they blur together into a singular entity: God. Or in polytheistic traditions, a pantheon of gods.
Even in hunter-gatherer societies, there’s evidence of proto-gods who mix with ancestors. In some Indigenous Australian languages, the word for “god” is the same as the word for “ancestor.” The great father god is basically the original ancestor of the tribe.
Öhman argues that this isn’t just some historical quirk - it’s how societies embody the collective wisdom of our past. In a totemic religion, everyone looks to a totem but the totem is a representation of old tribal leaders and every individual of the tribe. Christianity, in a way, is basically a totemic religion without a physical totem. Instead of carving a wooden idol, we project everything onto an abstract, all-seeing, all-knowing God.
Removing and re-inserting god in a secular society
As our society has become more secular, we’ve removed god from decision making. In the past, if someone asked: “Well why should King Philip rule?”, a sufficient answer would be “Because God said so.”
We’ve since removed god and taken the responsibility for decision-making onto human shoulders alone.
And now, we’re placing god back into the equation in the form of AI.
A new way to critique AI
Most AI criticism today focuses on clear, indisputable harms:
- Humans vs Machines – Ensuring AI aligns with human values, doesn’t wipe us out, etc.
- Humans vs Other Humans – AI as a tool of power and oppression
These critiques are important, but miss something deeper:
AI isn’t just another machine. It’s not just a tool of corporate and political power (though it is that, too). It’s something bigger—a new form of ancestral authority. AI serves as a unification of past human knowledge into a singular agency. AI is the wisdom of the past, reanimated.
The tyranny of the past
AI isn’t just about “machines taking over.” It’s about the past taking over.
AI models are trained on past data, which means every AI decision is, at some level, a reflection of historical human reality and judgements. When an AI outperforms a doctor in diagnosing diseases, it’s not “a machine” besting “a human”—it’s the amalgamated expertise of thousands of doctors from the past beating a single doctor in the present.
Should we let the past dictate the present? Even if it’s right?
AI as relationship counselor
Öhman suggests this scenario:
Imagine a couple consulting an ultra-powerful AI relationship counselor. They ask, *
“Based on all our data—biosignals, social media posts, personality traits—will we be happy together?”*
The AI, trained on billions of relationships, replies:
“98% chance this will end in disaster. You should break up now. Your chance of regretting it is close to zero.”
Technically, the AI is probably right. Statistically, it knows better than they do.
But something feels off. Most of us instinctively don’t want to live life this way. Why? If the AI is unbiased, fair, transparent, and accurate—what exactly is the problem?
The speaker argues that this gut feeling is valid. To articulate why, we need religious critique.
A human life worth living
The problem with a “perfect” god isn’t that it’s unlikely to exist — it’s that it would be undesirable.
Real freedom requires risk.
If we had a perfect, infallible authority guiding us, life wouldn’t really be ours.
Of course you want your wife to love you forever, but there is a risk she may not, and that makes you vulnerable in the relationship. It is a mistake to think you want to live in a world where it’s impossible for your spouse to leave you.
Yes, your ancestors, or AI, are probably right. The relationship probably will end in disaster. It is the wisdom of the internet accumulated, after all, and you are just one stupid, blinded person in love. But that is the beauty of being you. You can be in love and you can be stupid.
Love requires a leap, and romance requires an element of risk.
A human life worth living requires an element of risk.
AI as the Christian dream of perfection
For centuries, Christian thinkers saw technology as a way to restore human perfection—a return to our pre-Fall, Adamic state of immortality, omniscience, and flawless knowledge. Many Enlightenment-era scientists explicitly framed their work this way. Even Isaac Newton saw his discoveries as preparing humanity for the Second Coming.
AI is the culmination of that dream: the promise of perfect knowledge and freedom from error.
And just like religious critique has historically challenged divine authority, we need to challenge AI’s authority — not because it’s inherently bad, but because perfect knowledge is not necessarily desirable.
What Makes AI different from any other authority?
Isn’t this just like listening to parents, elders, or other experts? People have always relied on authority - what makes AI special?
When our parents say, “I don’t think this is a good idea,” you know it’s just one human opinion. If a group of experts say it, you know it’s a collective of individual perspectives. Even when you Google something, you’re still accessing multiple individual voices.
AI, on the other hand, doesn’t feel like a person or a group — it feels like a god. It’s presented as the distilled wisdom of humanity, an omniscient, impartial judgment. That illusion of divine intelligence makes it much harder to resist, and makes it different.
The big question
Öhman did not claim that we should ignore AI’s recommendations as a whole or reject expertise. His point, I believe, was that AI represents a new kind of authority, that we can critique from a religious standpoint.
So the real question isn’t just: “Is AI making the right decisions?”
It’s: “Even if AI is right, is a world dictated by AI still our world?”