Amy Welborn, Catholic author and blogger
Some of the most insightful reflections on artificial intelligence are not coming from Silicon Valley engineers, AI researchers, philosophers, or theologians. Sometimes they come from people who have spent their lives doing something that artificial intelligence is suddenly becoming extraordinarily good at doing: writing.
Amy Welborn is one of those people. A veteran Catholic author, blogger, and commentator, she belongs to a generation of Catholic writers who helped shape the contemporary Catholic literary and publishing landscape. She has written books on saints, apologetics, spirituality, and Catholic life, and has spent decades doing what serious writers do: researching, thinking, struggling with ideas, finding connections, and turning all of that into words that might illuminate something for someone else.
And now she is asking a disturbing question: What happens when a machine can produce, in minutes, what took her hours, days, or even years of intellectual work?
In two remarkable posts on her blog, Charlotte Was Both, published on October 2 and November 3, Welborn offers one of the most perceptive and refreshingly unpretentious accounts I have read of the crisis AI is creating for writers, particularly Catholic writers. Her reflections become even more compelling when read alongside a November 1 post documenting the flood of apparently AI-generated Catholic books appearing on Amazon.
What makes her observations particularly valuable is that she is neither a technological enthusiast nor an ideological opponent of AI. In fact, she doesn’t use it in her ordinary writing.
The invasion of the Catholic bookshelf
Welborn recently discovered a growing number of suspicious Catholic titles on Amazon, including books attributed to authors whose photographs and biographies appeared artificial or fabricated. Some of these books occupied surprisingly prominent positions in Amazon’s category rankings. One title she examined, Systematic Theology for Kids, had accumulated approximately 170 reviews whose authenticity she questioned.
Welborn is careful to acknowledge that she cannot prove every suspicious title was generated by AI, or that the sales and reviews are fraudulent. But the economic implications are obvious.
A publisher that hires an author to research the life of a saint, develop an engaging narrative for children, commission illustrations, edit the manuscript, and bring the finished product to market is suddenly competing with someone who can generate comparable-looking material in an afternoon.
Welborn even tested how quickly Claude could produce a saint’s biography for a particular age group, including questions and reflections. The exercise took approximately three minutes.
Imagine the implications for Catholic publishing, especially for smaller publishers whose already limited revenues depend on producing serious, trustworthy material.
But the problem extends beyond lost sales. It involves the possibility that Catholic families purchasing books about saints, theology, or spiritual formation will increasingly encounter material assembled by AI, packaged under invented author identities, and promoted through recommendation systems that may reward visibility rather than authenticity.
Welborn’s response to Catholic publishers is practical: stop depending so heavily on Amazon and rediscover the importance of building direct relationships with parishes, schools, dioceses, educators, and families.
That is an important suggestion, but I am not sure it can fly. Knowing some of the most prolific Catholic publishing companies, I can tell their marketing department continues to shrink because of the equally shrinking profit margin. They need to publish more for less revenue per publishing. Doesn’t that sound like a fertile ground for, well… AI?
Because, as Welborn herself discovers, the more troubling question is not whether AI can produce a cheap imitation of bad Catholic writing, but whether AI can produce a sufficiently convincing imitation of perfectly respectable Catholic writing.
When the machine produces something that is... actually good
Welborn freely admits that she does not use AI to write or even organize her ideas. She doesn’t rely on automated grammatical suggestions or ask a chatbot to solve intellectual problems for her. But curiosity eventually got the better of her.
In her October 2 post, she describes asking ChatGPT a highly specialized literary question she had previously researched extensively, presented publicly, and written about. The answer appeared within seconds.
To her astonishment, the machine reproduced essentially the same points she had reached after hours of research and reflection. Her reaction was not delight at having discovered a convenient research assistant. It was genuine uneasiness.
What, she wondered, was the point of having invested so much intellectual effort if the apparent result could be obtained almost instantly?
Of course, she understands that intellectual activity has intrinsic value. Researching, thinking, questioning, and writing change the person engaged in those activities. Learning is not merely the acquisition of information. But here is the important conclusion: the philosophical distinction doesn’t make the practical dilemma disappear.
Please go read an even more revealing experiment, described in her November 3 post, with an absurd assignment to Claude.
Welborn was shaken not because the result was brilliant, but because it wasn’t bad at all. Its quality was perfectly adequate for the genre. She wondered whether an ordinary reader, encountering it alongside a dozen human-written devotionals, would notice any meaningful difference.
In my humble opinion, this is precisely where so many discussions about AI go wrong. Because critics frequently insist that artificial intelligence cannot reproduce genuine human experience, spirituality, emotion, or creativity. That may be entirely correct.
But what happens when the reader simply wants an encouraging devotional reflection before breakfast and the AI-generated version satisfies that expectation?
The market does not necessarily reward the metaphysical distinction between human and artificial creativity. It rewards the product that meets the consumer’s expectations. And that’s a problem.
The experiment that frightened her most
Welborn’s third experiment was even more personal. She had accumulated approximately 30 blog posts on a subject she had explored over several years. Her intention was to organize these writings, revisit her recurring themes, and eventually develop more substantial articles.
She asked Claude to assemble the posts into a chronological document, preserving the text and hyperlinks while removing images.
The machine did the job, even producing a title page and table of contents. That was genuinely useful since an otherwise tedious organizational task had been simplified enormously.
So she went one step further: she asked Claude to write a publishable essay of approximately 1,500 words drawing on the principal ideas contained in those posts. Within five minutes, it had produced one.
Welborn read the opening and concluding paragraphs and was sufficiently disturbed that she stopped. She admits she had not read the entire essay, an important qualification. But what she saw appeared to capture her ideas accurately. Well, of course it did! The machine had been working with her own thoughts, arguments, and years of writing.
She followed with crucial questions, at the very crux of the issue: What if the resulting essay was not merely competent? What if it expressed those ideas just as effectively as she might have expressed them herself? Would that diminish the value of her years of work? Would it reveal that what she had regarded as distinctive creativity was, at least in part, a collection of patterns that could be reconstructed?
Welborn does not suggest that writers should surrender their craft and allow machines to do the thinking for them. Instead, she begins asking whether the arrival of AI might force writers to confront something they would otherwise prefer to ignore: how much of what we call creativity consists of producing familiar material in familiar ways for an audience that expects little more.
The uncomfortable truth about mediocre writing
One of Welborn’s strongest arguments draws on an essay by Jake Meador, editor of Mere Orthodoxy, examining how the internet transformed writing into a largely utilitarian activity. Her argument is compelling and elaborate, so I will let you go and read it.
But at the end of her musings, she ends up asking: How many of the countless daily Christian reflections published across websites, newsletters, and devotional books actually say something distinctive? How many are variations on the same reassuring formulas, familiar anecdotes, and predictable spiritual conclusions?
For a Catholic writer, this becomes an especially challenging examination of conscience: Are we writing because we genuinely have something to communicate? Because we have discovered something worth sharing? Because our encounter with truth has changed us and might change someone else?
Or because a publication schedule requires another article, an organization needs another newsletter, or a publisher needs another devotional?
And, perhaps most disturbing of all, if the latter is true, why should AI not be perfectly capable of producing the required material?
A question without an easy answer
The most remarkable aspect of Welborn’s reflections is her refusal to manufacture a comforting conclusion.
Her growing anxiety may not simply reflect fear of being replaced. It may reflect the unsettling realization that she needs to move beyond the habits and formulas within which she has become comfortable. In that sense, the arrival of AI might become a challenge rather than merely a threat.
Perhaps writers will have to become more courageous, more intellectually adventurous, more attentive to reality, and less willing to manufacture words simply because someone is prepared to pay for them. But will the market reward that effort?
Welborn does not know.
And I think her willingness to say so makes these posts considerably more valuable than another confident declaration that artificial intelligence will either save or destroy civilization.
Her central question is not a rhetorical device awaiting a predetermined answer. It is a genuine challenge: Where do we go from here?
The Catholic response cannot be another round of reassuring slogans
Readers of The Imago Project know that I have repeatedly argued for a Catholic engagement with artificial intelligence that avoids two equally inadequate responses: uncritical technological enthusiasm and reflexive rejection of a development we neither fully understand nor can realistically stop.
Of course we need the deepest possible anthropological, philosophical, and theological conversations about artificial intelligence: What exactly is intelligence? Is consciousness reducible to computation? Can a machine ever understand what it produces? What distinguishes human creativity from the statistical reconstruction of patterns? What does it mean that human beings are created in the image and likeness of God?
These are indispensable questions, but here is the uncomfortable fact: We can be absolutely correct about the unique dignity of the human person and still be completely unprepared for the economic and cultural consequences of artificial intelligence.
A Catholic writer can insist, correctly, that a machine cannot pray, experience grace, encounter God, or possess the spiritual interiority of a human being. But none of that prevents an AI system from producing a devotional reflection that thousands of Catholic readers find spiritually encouraging.
A philosopher can demonstrate convincingly that an LLM does not understand its own output in the way a human researcher understands an argument. But none of that prevents an editor from discovering that a research assignment that previously required an entire afternoon can now be completed in minutes.
In short, the transformative convenience of AI is not a theoretical proposition. It is something people are experiencing already. And convenience, when combined with adequate quality and dramatically reduced cost, has historically been among the most powerful forces reshaping human behavior.
Repeating that AI will never replace human beings does nothing to address the fact that it can replace many of the activities through which human beings earn a living, communicate their ideas, and exercise their talents.
Nor is it sufficient to compile increasingly long inventories of everything that can go wrong with AI, however real those dangers may be. Let me be clear: warnings are necessary, but they are not a strategy.
Why regulation alone will not save us
This is also why I remain deeply skeptical of the regulatory reflex exemplified by the European Union. The EU’s Artificial Intelligence Act, whose provisions are being implemented in stages, attempts to establish an extensive framework of obligations, risk classifications, transparency requirements, and restrictions on certain AI applications.
Some of these protections address legitimate concerns. Fraud, deception, violations of privacy, and the exploitation of vulnerable people are real dangers that no serious defender of technological progress should dismiss.
But the larger regulatory instinct strikes me as profoundly inadequate, because it seeks to respond to a technology of unprecedented adaptability, speed, and general applicability through legal structures developed for a world of comparatively stable products, identifiable producers, and predictable categories of risk.
Or to put it in practical terms, no regulatory classification can tell Amy Welborn how to preserve the value of her work when a machine can produce a plausible devotional in a minute; and no bureaucratic certification can restore the economic model of a Catholic publisher competing with an avalanche of cheap, automatically generated books.
I am not arguing for the absence of laws or accountability. I am arguing that an essentially defensive regulatory mindset cannot substitute for understanding a radically different technological reality.
The need for the Catholic imagination
I find Welborn’s reflections particularly compelling because they force us to confront the problem at the level where the transformation is actually taking place and not in an abstract debate about whether machines have souls.
The Church has lived through technological revolutions before. The printing press, industrialization, mass communication, electronic media, and the internet all transformed the conditions under which human beings lived, learned, worked, and communicated.
Catholics have not always responded wisely or quickly. Sometimes we have resisted innovations unnecessarily. Sometimes we have embraced them without sufficient discernment. But at our best, we have demonstrated an extraordinary ability to adapt new technologies to the service of truth, beauty, education, evangelization, and human dignity.
That tradition of intellectual creativity and practical adaptability is precisely what we need now.
We need Catholic entrepreneurs who can use AI to expand opportunities rather than merely eliminate jobs; working hand in hand with philosophers and theologians who can illuminate the human questions raised by AI without pretending that philosophical clarity automatically solves the practical problems.
Amy Welborn has confronted the possibility that a machine can reproduce much of what she has spent her life doing and has responded with intellectual honesty, discomfort, and without pointing toward a definitive solution.
And perhaps that is where the Catholic response to artificial intelligence must begin.