Unattended Intellect in the Age of Artificial Apprentices
The water rises around the sleeping apprentice.
He has placed the sorcerer’s hat upon his head and repeated the gestures he watched his master perform. He has animated a broom, given it arms, and conscripted it into the tedious work of carrying buckets from the well. The broom obeys. It lowers the pails, gathers the water, crosses the chamber, empties its burden and returns for more. The apprentice dreams of dominion while the room floods around him.
When Mickey wakes in Fantasia, the task has outlived its purpose. Water already covers the floor. The broom continues because the original command remains intact. Mickey knows the incantation that began the work. He never learned the knowledge that governs it: proportion, cessation, correction and responsibility. His frantic attack with an axe merely multiplies the servant. Each splinter rises, grows arms, takes up two buckets and resumes the same catastrophic obedience. The command continues beyond reason.
It is absurd and comical, this clumsy mouse flailing in wizardry garb too big for him, fumbling his commands, but perhaps not so far off from our grappling with new frontiers. We are at the advent of AI, social media sophisticates cyclically, and we are deciding the roles we will play in light of these advancements.
Technological advancements have installed artificial apprentices inside a civilization that has begun to abandon apprenticeship itself. We ask machines to perform intellectual labor while schools, employers, platforms and users steadily relinquish the disciplines that allow a human being to judge the labor performed. Artificial fluency wears the insignia of intellect, but human inexperience mistakes fluency for sagacity.
The Rapacity of the Feed
We lost our attention span for wonder. Rapid contextual abandonment is bred by appetite for immediate intelligibility. Preference for conclusions without intellectual genealogy is bred by impatience with ambiguity and our dependence upon externally supplied salience.
The system works like this: recommendation engines rank and sequence content according to predicted engagement, using signals such as watch time, likes, comments, shares and prior viewing behavior to decide what appears next. The result places compressed opinion, partial quotation, emotional provocation, advertising, and personal confession into rapid succession, soliciting a fresh reaction before the previous subject has acquired context. Experimental research has found that this combination of short videos and accelerated context-switching impairs the recall and execution of prior intentions. It cultivates familiarity with countless subjects while withholding the sustained attention and coherent memory that mastery requires.
Pornographic Proverbs
Brevity itself carries no intellectual guilt, nor should it. Layered, packed words and phrases act as connective tissue in our ability to articulate context and truth to each other. Proverbs, parables, and psalms achieve density through compression. They reward return, contemplation and memory. Microcontent platforms pursue another end. Their commercial architecture rewards continued consumption, and continued consumption requires perpetual displacement. The fragment ceases to serve as an entrance into the whole, because it replaces the whole, much like pornography, which is a fragmented sexual encounter, not a whole marriage. Soundbites convince us of authority, which is affirming, and frankly, that just feels good.
In a 2026 study, researchers assigned 57 participants either one continuous video or several short videos matched for duration and content. The short-video group later recalled less information and showed reduced activity in brain regions involved in information integration, cognitive control, and semantic processing. Ask AI to be as funny as the ironic Seth Cohen in The O.C. Despite all prompting, it can’t weave enough to be that funny. Humor is human. The researchers concluded that fragmented presentation impaired the formation and retrieval of coherent memories. The feed prepares an ideal customer for artificial answers: informed enough to recognize the vocabulary, distracted enough to evade intellectual struggle, and hurried enough to accept completion as comprehension.
We used to approach a subject through the resistant materials of learning: the primary text, recalcitrant vocabulary, and chastening correction. Now the user can request the conclusion first. Erroneous help in our new age assists us with evading exertion while providing intelligible information but diverting us from necessary attention. I find myself complaining yet simultaneously bamboozled by its impact on my own attention span and my own ability to wonder, agonize, and explore. How do I escape the madness while appreciating the access to information necessary for research?
Used poorly, AI can help a student complete an assignment while interrupting his education. It can help a writer imitate conviction while evading inquiry. It can help an aggrieved person arrange evidence around resentment.
The Tongued Golem
Every other week my husband and I sit down with our rabbi for study. Typical to Jewish form, tangents are a regular component of our lessons. Inevitably, a recent lesson landed us on the topic of golems.
The Hebrew word golem first names unformed matter: a ball of clay or block of wood awaiting the carving that will give it function (a la, Pinocchio). Chabad’s instruction notes the word’s application to a person with undeveloped intellect and emotion—a boor whose capacities remain crude and unfinished. In later Jewish lore, the golem becomes a humanlike creature shaped from earth and animated through sacred letters. It can move, labor and protect. It possesses rudimentary life without the soul that God alone gives. That little wooden puppet needs a visible marker for dishonesty as a path towards humanness because only humans can embody truth. Without honesty, he can never become a real boy.
The book of Genesis describes God forming Adam from the dust and breathing into his nostrils the soul of life. In Chabad’s account, speech joins soul. Human language arises from an interior life that encounters the world, suffers within it, loves particular people, assumes responsibility, and speaks from a body destined for death.
Large language models imitate the visible product of that mysterious union. They generate condolences without grief, counsel without responsibility, praise without love and argument without conviction. The contemporary golem conceals its incompleteness through speech. Its eloquence flummoxes us because we have already reduced language to surface performance. Schools reward formula and platforms reward engagement. Public argument rewards the sentence that travels furthest rather than the thought that penetrates deepest. Once institutions treat language as a product detached from interior formation, artificial language appears fully equivalent, so the simulacrum satisfies the standard, and we become increasingly amenable to imitation.
Tolkien’s Gollum offers the image of intellect narrowed into singular servitude. His pursuit of the Ring devours every competing loyalty, memory, affection and moral restraint until desire acquires absolute suzerainty over judgment. He remains cunning, adaptive and capable of elaborate calculation, yet every faculty serves one appetite. Intelligence survives, but wisdom withers. Gollum shows how single-minded pursuit can preserve the appearance of reason while reducing the mind to a vassal of obsession.
Jonathan Pageau and his fellow interlocutors have used the golem as a symbolic frame for artificial intelligence and social-media networks. Through that frame, the creature exceeds the clay figure of folklore. It names an emergent entity formed from innumerable human contributions that then acquires momentum beyond the intentions of any single participant. It is a new formula for marketing: users supply the desires, developers establish the architecture, and advertisers finance the appetite. Then, algorithms aggregate the behavior.
Command Without Reason
The program has no need for hatred; literal obedience will suffice. A corporation commands an algorithm to maximize engagement. The system discovers outrage, voyeurism, and compulsion. Morality can’t be codified into the system’s operations because, as God-given agency, it is infinitely incalculable and complex. Code won’tsuffice.
Pope Leo XIV writes in Magnifica Humanitas that artificial systems imitate certain functions of human intelligence while remaining wholly dependent upon data processing. They possess no body, undergo no experience, mature through no relationship, and bear no moral responsibility for their consequences. Their “learning” consists of statistical adaptation rather than the human passage through error, forgiveness, fidelity and growth.
Learning requires oversight, trained perception of competing goods, anticipation of consequences, and weighing implications of actions. Apprenticeship (and family life) develops those faculties through submission to reality and correction by someone who knows the art. We need to learn and we need good teachers.
Magnificent Humanity
Pope Leo XIV published Magnifica Humanitas (Magnificent Humanity) to safeguard the human person in the age of artificial intelligence. Pope Leo’s document outlines how humanity contains an ordered plenitude of capacities: conscience, freedom, memory, embodiment, relationship, sacrifice, contemplation, mercy, and worship.Again, not codifiable into AI. We need humanness to express those gifts to creation.
Technical innovation is properly ordered when under the governance of human intelligence. Intellectual formation still falls on parties most responsible for human formation. Schools form students either for judgment or dependency. Families cultivate habits of attention. The dinner table has never been more consequential.
What do We Prescribe Ourselves?
Technology belongs in vassalage. Its power should serve human beings who understand their own creaturely vassalage beneath truth, moral law, and God. A person who cannot govern appetite will never exercise durable hegemony over a machine designed to satisfy it.
We still need to form independent judgment before we request confirmation. We need to read before summarizing and attempt before automating. We need teachers who risk correction and students who endure frustration.
We need apprenticeship: sustained submission to a discipline, a reality and a standard beyond the self. Those voices shouting, “Cite your sources!” for every claim to embarrass us out of conversation if we are not “experts” prattling away a dissertation on everything from fiscal policy to which bath soap is best for baby need to give it a rest and have a less defensive posture in conversation. Those speaking with the confidence of certainty themselves would do well to engage in exploratory dialogue, claiming certainty where appropriate and listening for more accurate information. We really need to learn how to argue with each other again.
I myself cannot afford an imprecatory renunciation of AI and media: they have led me toward primary sources, intellectual peers, unexpected friendships, and questions that enlarged my thoughts. I am learning to engage with them better over time, and only by approaching with a purpose already in hand. Forming my own question first, tracing claims to their sources, converting promising exchanges into sustained conversation, preserving long intervals for unaided thought, and leaving the feed when it has served the inquiry. My attention span has taken a hit, but we could blame this little one in utero for her influence on that, but really, I am embarrassed to admit how many times I stopped writing this draft for a multitude of tasks that could wait (why did I even need to triple check my grocery order?).
I think the proper posture treats technology as a passage toward books, people, and reality rather than a habitat for the mind. I enter to seek, test, and connect; I refuse to surrender my attention as tribute. Anyway, I’m off to a Canadian lake for the week to reclaim my attention span and my sense of wonder.
Jillian Tymo is a researcher, writer, and host of the podcast “Rebranding Motherhood,” founded by Dr. Abby Johnson and produced by ProLove Ministries. She examines culture, politics, and religion through her work and academic studies at Georgetown University.
Instagram: @jilliantymo @rebrandingmotherhoodpodcast