The Vatican Has Some Thoughts on AI

Quick heads up: There are only four more days to register for my course, How to Write With AI. The four-week cohort-based class runs from February 13 through March 6 and includes:

• 4 live lectures and hands-on workshops
• A writing group overseen by an Every-trained editor
• Interviews with successful internet writers, including Every CEO and cofounder Dan Shipper
• Demos of cutting-edge products including Deep Research, Operator, and others
• 30 days of quick writing exercises
• Your own customized LLM prompt for improving your drafts
• A chance to share your writing with Every’s 93,000-plus readers

Over 90 students—including founders, aspiring writers, engineers, and professors—have already taken the course. Check out the website for more information and to enroll. Act quickly—registration will close on Sunday night:

---

AI is a scary technology. Not just because its power is mostly concentrated in the hands of residents of Palo Alto. Not because, if its promise is realized, it may make most of us unemployed. Not even because we are likely less than three years away from AI being able to replace the very column you are reading. (This is more of a personal fear and also related to my point on unemployment.)

AI is scary because it forces us to re-examine the big questions: What is intelligence, what is consciousness, or even, what does it mean to be human? It is thus unsurprising that the pope, something of a specialist on such existential questions, is weighing in.

The Vatican released a document last week called "ANTIQUA ET NOVA: Note on the Relationship Between Artificial Intelligence and Human Intelligence", in which they do their best to help resolve some of this dread. It may surprise you, but this isn’t the Catholic Church’s first foray into wrestling with the implications of AI. The Vatican has been working on this since at least 2016, when it hosted a panel at South by Southwest, and the pope has met with Meta chief Mark Zuckerberg and a variety of other tech luminaries over the last several years.

The depth of their effort—across 13,217 words and 215 footnotes—shows. Most of the work on the relation of the soul to AI has been laughably thin. This note is different. They lay out a theory grounded in Christian theology and classic philosophy that I found moving and, more importantly, helpful. There are some shortcomings, but overall, this document is profound. My goal today is not to argue for a particular faith tradition or theological view. However, I find that studying a faith that is very different from my own personal truth is useful as I consider the impact AI is having, and will have, on my own life and career.

What constitutes intelligence?

The crux of their argument is that there is a profound difference between what an LLM does and what human intellect is. An LLM does fancy math and generates an answer by finding the statistical distance between data points. Human brains may function similarly (there is still a lot we don’t understand about them), but in the Church’s view, that’s irrelevant. Rather than merely relate human capability to our success at finishing tasks, the Church argues that there are two key points of differentiation between human intelligence and large language models:

Rationality: Quoting Aristotle, the Church states that “all people by nature desire to know.” This desire to know, our innate hunger to grow in understanding, can be recognized in the complementary ideas of reason and intellect. We exercise intellect when we are able to intuitively grasp the truth. Reason or rational thought, on the other hand, is the “discursive, analytical process that leads to judgment.” In my experience, in Silicon Valley the idea of rationality is often confined to what can be rigorously structured, Excelified, and argued via a dashboard full of data.

The Church takes a more expansive view of rationality to include “all the capacities of the human person, including those related to ‘knowing and understanding, as well as those of willing, loving, choosing, and desiring.’” In the Church’s eyes, this pairing of intuition with reason amounts to the “image of God” in which all humanity has been created. The matter of defining human intelligence, then, is a matter of defining the nature of Divinity. 

Embodiment and relationality: I loved my recently born daughter as she progressed from an ideal hope to a positive pregnancy test. Just the idea of her existence was enough to get me excited. But when I held her on my chest, there was a profound shift in the intensity of my feelings for her. In the evolution from concept to physical reality, I understood my relationship to her so much more clearly.

The Vatican’s note argues this embodiment and relationship to physical experience is core to human differentiation versus LLMs. The “human person is simultaneously both material and spiritual.” Where our body and souls meet is what constitutes a human being. This point is crucial because the embodiment of our intelligences pairs our intellect with the senses, emotions, and relations to others. “Human intelligence is not an isolated faculty but is exercised in relationships, finding its fullest expression in dialogue, collaboration, and solidarity,” the Vatican writes.

Perhaps it’s a gross simplification, but “ANTIQUA ET NOVA” argues that human intellect is superior to that of LLMs because it can emotionally and physically intuit in ways that inference machines cannot. 

“At the heart of the Christian understanding of intelligence is the integration of truth into the moral and spiritual life of the person, guiding his or her actions in light of God’s goodness and truth. According to God’s plan, intelligence, in its fullest sense, also includes the ability to savor what is true, good, and beautiful.”

On a metaphysical level, this resonates. I do feel there is something profound in human nature and capability that LLMs are unable to do. On the other hand, accountants exist. Much of the intellectual labor that the human race is engaged in does not require embodiment or rationality, but rather the ability to manipulate data on a screen. Even if the Catholics are right, the majority of our labor today is menial, limp, and lame.

Applications and warnings

After spending 48 paragraphs arguing about the nature of intelligence and the implications of this definition, the document devotes the remaining 69 paragraphs (!) to specific applications of AI. It covers healthcare (AI shouldn’t be making care decisions), emotional relationships (AI shouldn’t be used as an emotional crutch), and warfare (AI shouldn’t be making any lethal decisions), among other domains.

What is lacking in the note, is, ya know, a solution to having people treat AI (and human beings) this way. While I find their definition of intelligence beautiful, meaningful, and true, their arguments have no teeth. There are simply no actionable suggestions to make their policies reality. 

There are vague talking points about how “the concentration of power over mainstream AI applications in the hands of a few powerful companies raises significant ethical concerns” (an objectively hilarious statement from one of the world’s most powerful and wealthy organizations). The authors never dive into specific issues like chip restrictions, open-source models, or anything that would actually make their meditations on AI actionable. 

Even if they exerted the full, holy might of the papacy, I very much doubt they could slow things down. American companies have announced an investment of $676 billion in data centers for AI this year. China is awakening with DeepSeek. The Saudis are firing capital everywhere with a $100 billion fund. No matter the strength of the argument or the spiritual need for us to reconsider what we are doing, the wheels of capitalism never stop turning. 

What scares me most about AI is that the Catholics are probably right—that human intelligence is unique and beautiful, and there is nothing we can do to slow down AI progress enough to protect that. Despite what they say, I’ll still use these tools because if I don’t, I won’t be able to provide for my daughter. 

---

Evan Armstrong is the lead writer for Every, where he writes the Napkin Math column. You can follow him on X at @itsurboyevan and on LinkedIn, and Every on X at @every and on LinkedIn.

Was this newsletter forwarded to you? Sign up to get it in your inbox.

We also build AI tools for readers like you. Automate repeat writing with Spiral. Organize files automatically with Sparkle. Write something great with Lex. Deliver yourself from email with Cora.

Get paid for sharing Every with your friends. Join our referral program.