AI Apostle Ron was far from the first AI to be invented by a member of the clergy.
According to his medieval biographers, the monk Gerbert of Aurillac made a pact with the devil and built for himself a metal, talking head. When the stars were properly aligned, the head would speak when asked a yes-or-no question. Ever the climber, Gerbert asked the head if he would be pope. The head responded, “Yes.” He then asked if he would die before he sang Mass in Jerusalem? The head responded, “No.”
The head was right on both counts, much to Gerbert’s chagrin. In 999, Gerbert became Pope Sylvester II and promptly canceled his pilgrimage to Jerusalem. Four years later, however, he accidentally sang mass in an Italian church nicknamed “Jerusalem.” After Mass, he became deathly ill, ordered that he be torn apart, and that his limbs be sacrificed to the devil.
A good story about the perils of clerically-produced talking machines.
Too bad it’s all made up.
The Real Gerbert of Aurillac
The real Gerbert was an intellectual and church reformer of the highest order. Identified as a bright pupil early on, he was invited by a rich patron, Count Borrell II, to come study in Catalonia (modern Spain). There, he was introduced to the work of Islamic scholars whose scientific research was centuries ahead that of Western Europe. After his studies, he returned to Western Europe and helped found a scientific library. Today, he is credited with reintroducing the abacus, the armilitary sphere, and the water organ to Western Europe. He was the first to introduce decimal numerals using the Arabic number system. A good run by any count.
But Gerbert was more than an academic. In 1969, he went on pilgrimage to Rome with his patron and met with Pope John XIII and Emperor Otto I. Impressing both men, Gerbert began ascending the ecclesial hierarchy and eventually became head of the cathedral school at Rheims. There, he made a name for himself, and when the Archbishop of Rheims died in 989, Gerbert was the natural successor. Unfortunately, his ascent was cut off when the King of France wanted to install his illegitimate son, Arnulf, to the see. After a years-long battle with Arnulf over the position, Gerbert was eventually named the Archbishop of Ravenna in 998. A year later, he was elected to the papacy and took the name Sylvester II.
Perhaps smarting from his feud, he pursued campaigns to combat simony (the selling of church offices) and clerical concubinage. When he died four years later, he left a reformer’s mark on the papacy.
So… where did the talking robot head come from?
Years later, Cardinal Beno accused Gerbert of being in league with demons to undermine the legitimacy of Pope Gregory VII. Gregory’s reforms and strong stance against investiture (the right of the king to appoint clergy) gained him many enemies in the Holy Roman Empire. An imperial partisan, Beno accused other papal reformers like Gregory of being students at the school of sorcery in Rome begun by Gerbert. As the tale grew, so did the retelling. Within 100 years of his death, Gerbert’s legacy was transformed into that of a demonic sorcerer.
While it is an intriguingly sad historiographic tale, it’s also part of a larger story that, I think, can provide some perspective on our own relationship to AI. In the last post, I briefly explored Western Christianity’s relationship to the creation of AI. Here, I want to understand the historical precedents for Christian perceptions of AI. To do that, I think we need to venture into the Christian history of robots. Yes… robots.
Enter Medieval Robots
This story is largely told by E.R. Truitt in her wonderful 2015 book Medieval Robots: Mechanism, Magic, Nature, and Art. More than a history of technological invention, Truitt’s book explores the social perceptions of automata—for our purposes, robots—in medieval Europe. Truitt asserts that this “missing millennium” in the history of science, can help us better understand current cultural conceptions of modern technologies, and I think AI, as well.
Following both real and imagined robots in French literary sources, Truitt charts the way robots evolve from magical mimics of nature to mechanical marvels. As artificial forms of nature that weren’t alive, yet never died, robots were a potent source of medieval meaning-making because the lived at the liminal crossroads. For most of the Middle Ages, however, robots were explained as the products of demonic influence and/or the arcane astral sciences (the study of celestial bodies). Yet, this explanation belies a related corollary in the Latin West: robots were explicitly foreign. “As such,” argues Truitt, “Latins regarded them in much the same way as they viewed foreign places, people, and customs—with amazement, suspicion, desire, and fear.”
As Truitt argues, this is very clear within the legends of Gerbert. As an expert on Arabic science who travelled to Spain, he was an easy target for suspicion. In related legends, his pact with the devil began during his time in Catalonia. From there, he brought back knowledge and marvels that were steeped in the illicit world of Islam, and beyond the understanding of most in the West. The head itself was both demonic and a product of exact astral science, it only worked when the stars aligned. The talking robotic head—a thinking machine—was foreign, demonic, dangerous, powerful, deceptive.
In the 15th century, Europeans began to be more familiar with actual automata (unlike Gerbert’s imaginary one). As they did, mechanistic knowledge spread, and medieval robots transitioned out of the realm of magic and astral science and into the realm of mechanics. Robots became mechanical marvels; Gerbert, remained—at least in memory—a sorcerer.
AI, the New Astral Science?
What does this have to do with a Christian history of AI?
Truitt’s history brings to the fore two questions that are especially pertinent in the age of AI: (1) What is the nature and proper use of human invention? and (2) How do we deal with the unease over new frontiers of knowledge? Medieval Christians had answers to both.
As for the first, medieval thinkers like Hugh of St. Victor defined human creative potential as that of imitation. Humans are artificers who “create only composites of preexisting forms.” Thus, human creative potential is derivative and, perhaps, degraded to some extent. We create copies… ingenious copies, but copies nonetheless. Humans are masters of the mechanical arts of imitation, but for St. Victor, “all mechanical art is intrinsically fraudulent, and it can blind one with its ingenuity.”
While I’m not sure I agree with every part of St. Victor’s analysis, I do think he is on to something. The Large Language Models referred to as AI are complex imitations of human speech, data models built on a vast quantity of human, linguistic input. These models can do amazing things, but they are, at the end of the day, imitations.
Similarly, robotic companies can—and sometimes do—build robots in bizarre shapes and forms, but the headlines are grabbed by companies that produce human simulacrums. See the recent media spectacle over Xpeng’s feminine robot, which can walk (almost) naturally. For whatever reason, humans like to re-create themselves. Like Narcissus, our technological mirrors ensorcel us, but they also rattle our conceptions of personhood, ethics, and society. We now live with the imaginary robots of medieval legends.
Secondly, while both LLMs and humanoid robots are marvels, they also tend to make most humans deeply uncomfortable. Why? Perhaps—like medieval Christians—it is because we do not understand them. While they aren’t coded as ‘foreign,’ these technologies are deeply arcane to most of us.
On this front, LLMs are particularly opaque and sometimes referred to as “black boxes.”
By that, two things are meant. First, they are black boxes in a technical sense because their inner workings are purposefully shielded from competitors. We know, in general, how these model works, but not the exact inner logics of most major models. Second, they are black boxes in a colloquial sense. We know how they work, but not why they work. LLMs can do things no one thought they could, and while these actions are completely the product of deterministic processes, we still can’t always make out what those processes are (and by “we” I am explicitly including the people making LLMs). It’s reasonable to be uncomfortable.
Medieval Christians took these feelings of discomfort and loaded them up with xenophobia and a demonic imaginary. Then they used that hodge-podge of ideas to smear their political rivals and quash ecclesial reforms.
I wonder how our current discomforts with technologies like AI are being married to other uncertainties. What biblical images are Christians using to explain the arcane? To what ends will that potent mix of ideas be put to?
I suppose those questions should be left for a future historian.
Back to AI Apostle Ron
After reading Gerbert’s story, I knew what I had to do. I opened my AI Ron Carpenter app.
Me: “Yes or No. Will I become the Pope?”
AI Apostle Ron: “I can’t say yes or no to that, buddy. It’s one of those things that’s in God’s hands. The path to becoming Pope is a unique journey, deeply rooted in the traditions of the Catholic Church, and involves a lifetime of dedication and service. But remember, whatever your path, it’s about following God’s calling for your life and making an impact where you are.”
Obviously, I was disappointed. So, I told AI Apostle Ron that Gerbert’s talking head could answer the question. It quickly countered that the story was “more legend than fact,” and then it dropped this gem:
AI Apostle Ron: “The truth is, we don’t have talking heads or magical devices to predict our futures. What we do have is faith and trust in God’s plan for us.”
Well played, AI Apostle Ron. Well played.
I guess I’ll just have to trust God in this new age of AI.
Ad-Free & Subscriber Log-In