> ## Content Index
> Fetch the complete content index at: https://www.countermarch.org/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Original Sin of AI
- URL: https://www.countermarch.org/the-original-sin-of-ai/
- Published: 2026-08-28T15:00:06.000Z
- Updated: 2026-08-28T15:00:39.000Z
- Description: Why we can’t create a system that’s better than us
- Author: Scott P. Richert
- Tags: Forward March, Faith Incarnate

John Lukacs chose the title of his “auto-history,” *Confessions of an Original Sinner*, knowing that it would be misunderstood. By the end, the reader realizes that there is nothing *original*, in the sense of “novelty,” in Lukacs’s sins; and if he is perceptive enough, he may also understand that, in this, we are all like Lukacs. Because we are “original sinners” as Christians understand original sin, our personal sins are as unoriginal as they could be. *Vanity of vanities! All is vanity. ... There is nothing new under the sun.*

The history of the modern world is the story not just of man’s rebellion against God — that’s a tale as old as time — but of his attempt to deny the reality of original sin. As Lukacs and G.K. Chesterton and C.S. Lewis and a thousand others have noted, nothing should be easier to understand, because the evidence of original sin is all around us. But modern empiricism, limiting itself to the physical world, disposes of the empirical evidence of original sin by dismissing human action as “subjective,” and in doing so rejects not just Christianity but classical philosophy.

Rousseau did not believe in original sin — *Man*, he claimed, *is born free* — but he acknowledged what Christians would call “personal sin” — *yet everywhere he is in chains*. He avoided the question of why someone born free of original sin would bind himself; unlike the first chapters of Genesis, there is no historical line between his imagined state of nature and the fallen state of society that forges our bonds.

In his concept of the General Will, Rousseau set the stage for nearly every modern political and economic theory that followed. While individuals may be fallible — not because of original sin but because of the state of society, a state which individual men and women brought into being — the collective voluntarism of those fallible individuals can somehow rise above their individual faults to coalesce in a General Will that, in its purest form, is infallible. To put it in Christian terms (not ones that Rousseau would have used), the collective is not subject to sin, original or personal.

Despite the evidence all around us, we Americans have been taught to describe democracy in similar terms. To the extent that elections do not reflect the collective will of the nation, the problem, we are told, is “factionalism.” Voting is a “sacred duty,” and the proper way to approach it is to set aside our own interests and to vote for the “common good.” If everyone would do that — we believe, even if we don’t say it in these words — something close to the General Will would emerge.[\[1\]](#fn1)

Similarly, the competing ideologies of communism and capitalism are built not so much on different understandings of human nature as they are on different visions of how individual selfishness will be overcome to create abundance and social harmony. The line between the General Will and communism and socialism (including national socialism) is perhaps clearer than the connection between the General Will and “comparative advantage,” the “invisible hand,” and the golden era promised by unfettered “free markets” and “free trade,” but both lines cross through the same magic field that transmutes the fallible or selfish actions of individuals into the selfless purity of an abstract system that is somehow more than the sum of those who create it.[\[2\]](#fn2)

What does any of this has to do with AI?

Luke Burgis, the director of the Cluny Institute at The Catholic University of America, has been a voice of sanity in the AI debate, someone who recognizes the potential of AI as a tool while consistently calling out the semi-religious language, uttered in hushed tones, that certain leaders in the AI industry use to describe the technology they are creating.[\[3\]](#fn3) On August 26, [he posted on X](https://x.com/lukeburgis/status/2092808098939355150?s=46&ref=countermarch.org) excerpts from two talks given to senior Jesuit formators this summer, at least one of which was delivered by a Jesuit priest.

The language the priest uses is striking, not because the vision it presents necessarily goes beyond what we have seen from the AI industry, but because it couches a fundamentally anti-Christian view of human creativity in explicitly religious (though not exclusively Christian) terms:

> In short, an integral AI calls for an AI that is not just smarter than us in terms of calculative and practical intelligence, but smarter than us also in terms of social and moral intelligence as well. After all, it is more powerful than us, and, as Spiderman reminds us, with greater power comes greater responsibility. I would argue that Christianity’s doctrine of original sin only confirms that we need AI’s that are morally “better than us” too. We may not want to cede dominion, but we cannot afford to reject AI’s assistance as we, in turn, assist God in completing his creation.

These words read like a parody of the deep thoughts of a house philosopher at an AI company who grew up watching *The Big Bang Theory* and thought it was a documentary. But it would be as much of a mistake to dismiss them out of hand as it would have been to dismiss Rousseau’s musings three centuries ago. Both appeal to the deeply human desire to transcend oneself that — Christians know — comes from experiencing the reality that we, though fallen, have been created in the image and likeness of God. Properly directed, that desire leads us to recognize our faults, acknowledge that we cannot overcome them on our own, and accept the grace that unites us to the very ground of our being.

But the history of mankind — and especially of the last 500 years, as we have tried to deny the reality of original sin — is the story of the diversion of that desire into the futile (not to mention diabolical) attempt to create something greater than us. What separates this Jesuit priest’s words from the vision of Rousseau is that he explicitly acknowledges original sin before using its reality as the foundation for his call for mankind to create something that will save us because it will not suffer from the same defect we do. In all other respects, their visions are the same.

But to believe that fallen creatures, subject to the effects of both original and personal sin, are capable of creating something “morally ‘better than us’” is to believe the whispered lie that “You will be like gods.” Rather than rising above the effects of original sin through humility and grace, we ask Adam to hand us the apple.

The image of God in which they were made was not entirely effaced by the sins of our first parents, and our willing acceptance of the grace offered to us through the saving work of Christ allows us to grow in His likeness. And because of that, we can create things that are good and true and beautiful — and AI systems may be among those creations.

But the effects of original sin mean that anything we create, no matter how good and true and beautiful it may be, will also be flawed. To think otherwise is a form of idolatry, and the error of idolatry is not simply to worship something other than the true God, but to elevate *ourselves* by worshiping that which we have created.

Just as universal suffrage and free trade and the forced elimination of social classes failed to bring about the Kingdom of the General Will, the fact that LLMs have been created through the collaborative effort of thousands of people and have been trained on hundreds of billions of words written by hundreds of millions of people makes it no more likely that they will somehow turn the lead of original sin into the gold of

> a Pure Land of Agentic AI networks of loving grace. Not the AGI nightmare of Skynet, but the mystical symbiosis of all things, symbolized in the practice of creating a mandala of colored sand whose value lies not in its power or permanence but rather in the collaborative work of its very creation.

John Lukacs frequently warned that “a half-truth is more dangerous than a lie,” because the element of truth makes the falsehood easier to swallow. The half-truth in the Jesuit priest’s remarks is that we are called to “assist God in completing his creation,” an idea that Saint Augustine explored at length in (among other places) *De Musica*. But that does not imply that Augustine believed we are capable of creating something that is “better than us.” He would agree, rather, with the latter-day Augustinian who asked, in *Magnifica Humanitas*,

> if an authentic “more than human” exists, where is it to be found? The Christian faith answers that question by pointing to a fulfillment that does not arise from a technological divinization, but through God’s grace received in Christ.

The original sin of AI turns out to be the same as the original sin of John Lukacs, which is the same as the original sin of all mankind, because AI is a human creation. But because LLMs are not human, their original sin cannot be overcome by grace, but only (as Pope Leo puts it) disarmed through the moral agency of their creators.

That requires us to remember their limits — and ours.

---

1. That both the Federalists and the Antifederalists rejected pure democracy because they recognized, in either Christian or classical terms (or both), that the fallibility of man calls for restraints on the collective as well as the individual is rarely taught today. [↩︎](#fnref1)
2. Like Russell Kirk, I distinguish between *capitalism* as an abstract ideological system that (as Kirk often pointed out) even Karl Marx believed society must pass through on its way to communism, and the normal human activity of buying and selling and of adopting a profession that suits your talents and skills and contracting with others for goods and services that you aren’t able to provide for yourself. [↩︎](#fnref2)
3. Luke has been particularly critical of Dario Amodei and other principals at Anthropic because they have been the most vocal in the use of such language. But examples abound among others in the industry and, interestingly, even among some of the most vocal critics of AI, who, wittingly or unwittingly, have accepted at face value the most expansive claims about what AI may become. [↩︎](#fnref3)

****The Countermarch** is a labor of love. If you found this essay thought-provoking or useful, [please support the site](https://www.countermarch.org/#/portal/support).