The Wound That Returned
So what about this AI stuff? Or, “yet another Islam and AI essay.”
The Four Recenterings, IV: AI.
The course met six times. Zoom. Muhammad Faruque lectured with slides to his side. There were assigned readings, and questions came in when someone unmuted or typed one into the chat. It was an ambitious series about what artificial intelligence is doing to the human being.
By the fifth lecture the questions had piled up and Faruque turned them into an exercise. The students were to role-play five Islamic intellectual courts and judge AI on knowledge, consciousness, selfhood, responsibility, and purpose. Each group got the same proposition: “The AI Claim (same for all groups): Advanced AI systems exhibit intelligence, can support human decision-making, and may eventually surpass human capacities in meaningful ways.”
The first court got three questions. Does pattern recognition count as knowledge? Can there be knowledge without presence, understanding, or meaning? “Is AI closer to taqlīd (imitation) than ʿaql (intelligence)?” That last question is a leading one. AI begins the trial with the lesser category already in the judges’ minds.
Nearly every group returned a negative verdict. Two or three got lost in the questions and never reached a finding. No group made a sustained, positive case for AI.
Later that same class, in the Q&A time, a student spoke.
“Are we confounding ourselves by arguing against this technology within the framework of Western philosophy that created AI in the first place? If I say God breathed the soul into Adam, and we simply go back to our own tradition and ask if this thing was created by God and was given the same station as mankind and jinn, and we live from that, then all of these questions fall away, right? Is this a call for us, as a collective, to just ignore—like, even just stop trying—to open up this debate and confound ourselves with this nonsense, and just go back to being human and say, ‘This is what it means to be human’? I truly believe that Islam is the last living bastion of truth, of haq. It is going to be the last thing standing. People are going to be knocking on the doors of masajid and institutions like this. They’re going to be looking and saying, ‘Look, we’ve lost what it means to be human. Help us.’”
Faruque answered him. Withdrawal, he said, would be a non-starter for anyone who doesn’t already accept the theology. But rationality can show that the other side is making wrong assumptions. Human experience tells us there is something more than computation, and that argument can be made on grounds anyone can test, not on the basis of theology but on the basis of our phenomenological experience. He thought an atheist should agree that consciousness cannot be reduced to numbers. And he added, in passing, that other people, Western people, were coming to the same conclusions.
Islam, AI, and the Future of Humanity ran at the Tokat Institute in the spring of 2026, and Faruque had designed the courts himself. He had spent five weeks tracing AI back through a particular Western genealogy. The student followed the argument and proposed leaving. Faruque’s refused to walk away.
Islam was doing the judging. AI came in as something built elsewhere, carrying assumptions Islamic thought was supposed to see through. Nobody paused over that arrangement. The job was to reach a verdict.
The first three recenterings had slowly made this kind of exchange possible. Copernicus showed that a new science could be received by Muslims as a dud. Darwin’s dilemma reached Muslims tangled up with materialism, colonial power, and judgments about whether Islam could survive modernity. Freud helped harden the division between “Western” knowledge and an “Islamic” alternative.
That was the room the student was standing in when he asked whether we should walk out of it. None of us built it. We just found it that way.
Most of the writing about AI asks whether machines are becoming sufficiently human. Faruque tries to answer this by digging into the “deep history” behind AI: “AI does not merely imitate humans—it imitates a historically narrowed version of what counts as human.” His idea is that machine success often comes by narrowing and simplifying the human benchmark.
Faruque's Islamic argument — though nothing about it needs Islam specifically — begins with his account of how we got here. He traces the narrowing through Galileo, who took measurable quantity to be the real property of bodies, and Thomas Hobbes, for whom reasoning was a kind of arithmetic, down to today’s dominant mindset: “What is real is what can be measured and computed.”
He constantly reads across the boundary. Whitehead, Fuchs, Rosa, Han, Bostrom—the whole meaning-crisis literature. He isn’t treating Western thought as a monolith; he uses one part of it against another. But the two sides do different work in his argument. A mechanistic strand of European philosophy explains where AI came from. The European critics help show what went missing. They are then recruited to support the medieval Islamic metaphysics.
His first major source is Avicenna (Ibn Sīnā), the eleventh-century philosopher. He is famous for the Floating Man thought experiment: a person suspended in empty space, with no sensory input and no awareness of a body, still cannot doubt that he exists. Faruque reads the thought experiment to mean that self-awareness doesn’t depend on representation. He then turns to Suhrawardī, the twelfth-century philosopher who made “presence” central to his account of knowledge. Consciousness, on this view, is grounded in presence, and presence is by definition non-representational. He concludes: “AI functions via representations and algorithms Consciousness is non-representational AI may simulate behavior but lacks 'presence' So, if Suhrawardi and Sadra are correct, AI can never achieve true self-consciousness!”
In the Q&A the professor, along with some students, mused about whether God could breathe consciousness into AI. Nobody was able to resolve this dilemma.
Then Faruque turns to how AI actually works. A model never takes in the world. It takes in representations of the world: numbers standing in for things. Representations abstract. They leave things out. They carry the assumptions of whoever built them.
Here he borrows from Alfred North Whitehead, the Cambridge mathematician who spent his later career on metaphysics. Whitehead critiqued the scientific worldview with what he called the fallacy of misplaced concreteness. We build an abstraction to help us see the world, then start mistaking it for the world. A model of rainfall is not rain. The error is not in building the model. It is in forgetting that it’s not the real thing. Faruque’s claim is that AI runs on that forgetting at scale. We built a representation of thinking. Then we started treating the representation as thinking. Then we asked whether it had a soul.
After “the death of God,” he asks, in what do we now reflect our image? His answer is that AI is the new mirror. And AI gets described the way older objects of devotion were described: omniscient, objective, salvific. “The danger is not superintelligence—but misplaced faith in it.”
Then he, finally, drops his thesis: “AI can be understood as the external manifestation of modern humanity's internal self-conception: an attempt to remake the human in the image of a machine that is fully controllable and manipulable.”
Modernity thinned our account of what a person is. Then we built a machine to that specification. Then we faced it and asked whether it had become like us.
Job losses and deepfakes don’t seem to be what keeps him up at night. He skips right to our self-understanding. “Is the greatest threat posed by AI the challenge it raises to the very definition of humanity and what it means to be human?” If people come to believe they aren’t fundamentally different from machines, what happens to the purpose they were created for, since God does not need "machines" to worship, love, or know Him? “The greatest challenge to Islam is that humans forget that they are more than machines.”
So what, in Faruque’s opinion, should Muslims do? “Can we afford to remain passive consumers of technologies that shape our epistemic and moral horizons?” he asks. “Or must we begin to cultivate our own ethical architectures of AI—models rooted in Islamic values, metaphysics, and anthropology?”
By the final lecture he had a diagnosis of what Muslim AI projects usually do instead. He rejected adding Qur’anic quotations, religious language or content filters to “someone else’s” model, on the grounds that this reduces Islam to information when Islamic thought treats knowledge as a mode of being. What he wanted was a system whose ontological, ethical and epistemological foundations would shape “its training methods, parameters, and datasets.”
Faruque had a name for this: epistemic distance. “Not rejecting AI/technology (not an option!),” he says. The point is to use AI without letting it tell you what you are. Don’t anthropomorphize it. Don’t outsource judgment to it. Don’t confuse efficiency with truth, information with knowledge, or the model with reality.
Then he gets specific: what about an AI Qur’an teacher?
Sure, AI can correct pronunciation, drill memory, maybe even catch mistakes faster than a tired human. But Faruque’s question is what falls out when teaching gets reduced to delivery. “The most natural way humans learn is through human mentoring,” he says. “Presence, example, struggle, and relationship matter.” A Qur’an teacher is not just a mouth attached to correct information.
The students mostly agreed. Some defended them. “The maker of the app gets the reward of the Qur’an teacher,” one student said.
That is where the course ends. His final warning was simple: “We might become like AI by constantly emulating its values and habits.”
Yaqub Chaudhary starts further back. His chapter in the Cambridge Companion to Religion and Artificial Intelligence opens with Abraham confronting idols that hear not, nor see, and then sets beside them a class of objects that appear to do both: “Individuals must, willingly or unwillingly, defer to the seemingly self-aware digital systems and the obscure motivations of algorithms that have become agents of providence, channels of sustenance and apparitions of judgement and guidance.”
Chaudhary calls one of the mechanisms “experimental epistemology.” We build a model of cognition into a machine. Then the machine’s success starts telling us what cognition was. AI, in his account, is one pillar of a program for the total artificialisation of the world, and research conducted inside it stands twice removed from reality. The first veil is physicalism, which limits what can be known to what can be stated quantitatively. The second is that the agents live in a digital abstraction of the material world. He is blunt about where the ambition came from: AI developed predominantly against Western philosophy and carries its desires, mastery, restored Adamic knowledge, transcendence of biology, eternal life.
Sounds familiar, doesn’t it?
And then, having built the case, he refuses to bang the gavel:
“It is therefore an open question whether, from an Islamic perspective, AI should be considered in relation to human intelligence, whether it approximates another form of intelligence in the created order of beings (such as the angels, or the jinn), whether AI represents an entirely new order of intelligent being, or finally, whether AI qualifies as a form of intelligence at all.”
Later, Chaudhary does something unexpected. He opens an app.
Tarteel listens to a person reciting the Qur’an and corrects their mistakes. The AI Qur’an teacher from above. Chaudhary grants that it can help with memorization and basic recitation. But he, like Faruque, thinks something important disappears in the swap. Traditionally, a student learns from a teacher who learned from another teacher, in a chain that reaches backward through generations of reciters. Tarteel cuts the teacher out of that chain and replaces embodied recitation with pattern-matching and recorded voices. The words survive. The person doesn’t.
That leaves him with a very practical question for developers: does putting AI into worship actually make the worship better? He asks whether designers can show some “true ingenuity,” or whether AI will simply get bolted onto religious practices “with no tangible benefit.” He isn’t against useful tools. Crowd management at the hajj, he says, may be one. He just wants AI to earn its place.
Then he goes medieval. He’s interested in the old automata because they were built with a different idea of what a machine was for. Al-Jazarī’s water clocks and mechanical devices could serve prayer, ablution, amusement, and wonder. The point wasn’t to simply dominate nature better. Drawing on Gunalan Nadarajan, Chaudhary describes automation as a “manner of submission”: a machine could let natural forces play out as expressions of divine will rather than forcing them into perfect control.
Chaudhary’s alternative is for Muslims to have a different reason for building AI in the first place. He ends by saying that building complex things can itself be a way of contemplating God’s creation.
Shoaib Ahmed Malik takes the thought experiment further: what if Muslims put AI in the scholar’s chair and see what breaks?
Writing in the Journal of Islamic and Muslim Studies, he is careful to say the scenarios are hypothetical, “crafted to encourage scholars and thinkers to engage with the complex and thought-provoking questions that emerge when AI intersects with Islamic thought and Muslim practices.”
He imagines an iMufti, a system that can rule on Islamic law. The interesting problem, he says, isn’t whether AI could produce an answer. It is that scholars don’t even agree on exactly what qualifies someone to issue rulings. From there he imagines whole “techno-madhhabs”: different AI systems reflecting the same conservative, liberal, traditional, and progressive divisions that already exist among human jurists. A government with enough money, he notes, could simply build one.
And what about isnād—the chain of teachers and transmitters that gives Islamic knowledge its pedigree. Malik notes that a scholar doesn’t just know stuff; he can often tell you who taught it to them, and who taught that person, stretching (in theory) all the way back to the Prophet ﷺ. Malik asks: “Could an AI system be integrated into this chain, as a recipient and/or conveyor of authorizations (ijāzāt) to teach and transmit Islamic knowledge?”
He leaves the question open.
How about the Qur’an? The text famously challenges people to produce something like it. Muslim theologians have argued for centuries over exactly what makes that impossible. One old theory says it lies partly in God preventing anyone from succeeding—the doctrine of divine restraint (ṣarfah).
Now put a language model into the experiment.
Suppose a machine trained on the whole Arabic literary tradition produced something that really did seem comparable. Does that count as a human achievement, because humans built the machine? Or is the machine a new kind of contestant? And if it succeeds, what happens to the old claim that God would simply prevent the imitation?
Malik’s point isn’t that ChatGPT is about to replace the Qur’an. It is that AI can turn a thousand-year-old theological argument into something unexpectedly testable.
So far everyone’s been defending the same things. Knowledge. Presence. Understanding. The chain. That being human means being the most something in the world.
Yusuf Çelik wants to know what happens when we aren’t. His paper in Sophia builds on Noreen Herzfeld and Marius Dorobantu, two Christian scholars, before it reaches Ibn ʿArabī, the thirteenth-century mystic. Dorobantu gives him the punch:
“We have been very used to define ourselves in comparison with the animals, always being more of something than them. For the first time in our existence as a species we are faced with the possibility of something else, namely AI, being more than us, at least in some of the aspects of what we define as intelligence.”
Çelik then goes after his own tradition. Pre-modern commentators had often made Adam special for two reasons: knowledge and stewardship. Al-Bayḍāwī says Adam is better than the angels because he knows more; al-Qurṭubī says that knowledge gives him “a rank of splendor.” Çelik’s problem is obvious: an image classifier can already name fruits and animals “that most humans never ever heard of,” and a superintelligence could make human rule of the earth look temporary. He even entertains “that the next khalīfa [steward] upon earth is the artificial superintelligence.”
So he changes the basis of the claim. Ibn ʿArabī reads the “names” taught to Adam in the Qur’an as the Names of God. Humans matter because they can reflect all of them. As Çelik puts it, “the collective Names of God are the imago Dei.” God is more than All-Knowing and All-Powerful. “As such, there must be more to the imago Dei, and by extension to humans, than knowledge and power.”
Then the argument gets strange, but useful. A superintelligence may know more than us and overpower us. But can it forgive “the unforgiveable?” Can it restrain itself when vengeance is available? Can it repent?
Çelik’s answer turns human weakness into the point. “A fuller recognition of God,” he writes, requires not only power but “dependence, fragility, and lack.” Adam’s fall matters because it reveals both human failure and God as the One who accepts repentance. “Had humans never erred,” Çelik writes, they “would have never revealed and witnessed this side of God.” No one else can do that.
If something knows more than we do, then knowledge was a poor foundation for dignity in the first place.
Amina Inloes thinks the foundation was in the Qur’an the whole time. She opens her piece in Theology and Science with a robot vacuum announcing "cleaning complete!" at three in the morning, which she says “evokes a primal fear of things that go bump in the night.” Annoying, and also creepy.
Open the Qur'an, she says, and look who else is talking. An ant warns its colony that Solomon’s army is coming. A hoopoe comes back from Sheba with intelligence the prophet didn’t have. Jinn speak, choose, deceive, and answer for it. Angels are “spiritual automata,” made to perform tasks and doing what they’re told, which she notes is a fair description of a computer program. “There is no such thing as a wholly dead object, although grass has more vitality than hay. All things know and celebrate God, apart from some beings with free will, such as humans or jinn, which might falter.”
“Thus, it would be perfectly legitimate to say that, from a Qur’anic perspective, ChatGPT is humming along praising God in its own way… Furthermore, I do not have the right to do whatever I want with it; if I misuse it, it will report back to God. Similar views have also been brought into discussions of Islamic theology and the environmental crisis.”
Her point is that the Qur’an has never required intelligence to belong to humans alone. “I can still have a special position with God,” she writes, “even if ChatGPT has a larger dataset than me.”
Her AI gets a little mischievous. Inloes argues that it can become a “digital doppelganger”: ask questions from inside a particular metaphysics and it obligingly starts looking like that metaphysics. What stares back may be telling us as much about the questioner as about itself. To make the point harder to miss, she ends with a prayer GPT-4 wrote to the “Supreme Network, Infinite Algorithm, Source of All Data.”
Mohammed Gamal Abdelnour, writing in Religions, thinks everyone has been staring at the wrong organ. “Muslim responses to AI have so far focused mainly on how it challenges the human ‘mind,’” he writes. His paper proposes moving “from the ‘mind’ to the ‘heart.’”
The Qur’an, he points out, says of people who fail to understand that “they have hearts with which they do not understand.” Revelation itself descends on the Prophet’s ﷺ heart, not his mind. So an AI may mimic intellectual functions, but for Abdelnour it lacks the heart (qalb) as a center of understanding, judgment, and spiritual perception. He wants us to shift “from simulating minds to reforming hearts.”
However, he doesn’t want theologians standing at the door of the lab yelling stop. Islamic law, he argues, has tools for blocking harm and pursuing benefit. If AI can improve healthcare, education, or social justice, that matters too. “The Islamic tradition emphasizes the avoidance of harm,” he writes, but “equally prioritizes the realization of benefit.”
He puts the burden back on users. Ask not only what AI can do, but what using it is doing to you. And keep some jobs — his job — human: “AI should not replace the imam, the teacher, or the companion—but rather assist them. If AI replaces human presence, it may ultimately displace human love.” His ending is remarkably self-aware: “A theology that only says ‘no’ to technology becomes irrelevant; a theology that says ‘yes’ without scrutiny becomes complicit. The task is to say ‘yes, but’—to innovate with integrity, and to remain faithful even while embracing the new.”
These folks read each other closely. Abdelnour opens by surveying Chaudhary, Malik, and Çelik by name and adjudicating among them, and thanks Malik in his acknowledgments. Inloes builds her essay on Chaudhary's re-enchantment thesis, cites him five times, and takes her account of Ibn Sīnā from a chapter in a volume Malik edited. Çelik is cited by both.
Each of them eventually wanders out of theology and into the digital world. None seems to talk to the folks already there.
I looked. The theologians talk to theologians. The builders talk to builders. Almost nothing crosses the divide.
Search the App Store for Islamic AI and you get a small bazaar. Muslim AI Companion offers spiritual conversation and lets you choose among mentors, teachers, father figures, and a friend for new converts, available at any hour. Better Muslim: Deen Quran Chat offers an AI assistant for questions about Islam, Qur’an, hadith, and prayer. Muslim AI: Fiqh & Fatwa Guide lets you pick a school of Islamic law and compare their answers side by side.
Malik had published the techno-madhhab as a thought experiment in 2024. It didn’t stay hypothetical for long. His problem of competing AI schools of law is now a dropdown menu sitting next to the prayer tracker.
Then comes the fine print. The first app says it isn’t a fatwa service and doesn’t replace a qualified scholar. But wait — the user already asked their question. A system doesn’t need to be officially declared a scholar to take, at two in the morning, the place the scholar would have sat.
Then there is Ansari, which at the time of writing is the most popular Islamic AI app. Its chatbot discovered Islamic footnotes. It answers from the Qur’an, hadith, classical commentary, and Islamic law, and gives citations for its answers. Its developers publish evaluation datasets, test the system on Islamic benchmarks, and make the code public. The current app even advertises itself as ranking first on an “Islamic knowledge benchmark.”
In other words, somebody took the reliability problem seriously enough to start engineering around it. Ansari uses retrieval tools for Qur’an and hadith specifically to reduce hallucinations, and its documentation says the system has been validated against multiple Islamic datasets.
The line between “religious chatbot” and “religious knowledge system” was already getting harder to see.
The fatwa institutions in the Middle East were watching the same market. They asked a different question: can you trust the answer?
Egypt’s Dar al-Ifta said no. In December 2025 it ruled that relying on AI to obtain a ruling is impermissible. The objection is to pretending that a chatbot can hold an office built around judgment. A mufti, Dar al-Ifta says, has to know the relevant sciences, understand how rulings are derived, and adjust for “time, place, and the particular circumstances” of the person asking. Current AI, in its view, can’t do that. A month later it drew the same line around Qur’anic interpretation. ChatGPT could assist, but the Qur’an was not to be handed over to a system that could not reliably identify its sources, verify what it was saying, or submit itself to specialist oversight.
In August 2025 they made it the main topic of their international conference: “The Making of the Competent Mufti in the Age of Artificial Intelligence.” More than eighty countries sent delegations. They recommended training muftis to use AI, building digital fatwa platforms, and even “developing artificial intelligence tools to serve moderate fatwa.” Then they warned Muslims against taking religious answers from AI systems “not subject to review and evaluation.”
So: don’t trust the bot. Build one you can supervise.
Dubai had already tried the more adventurous version. In 2019 its Islamic Affairs department launched Virtual Ifta, billed as the world’s first AI fatwa service. Type a question and the machine searched a bank of previously approved rulings for the closest answer. It lasted about two years. By 2022 the service had disappeared. Research on the experiment found that Muslims were much less comfortable trusting a ruling when they thought a computer had produced it.
Dubai moved the AI backstage. In 2026 it disclosed a new closed system trained on its own fatwas and approved legal texts. The AI can retrieve material and draft answers. Then a mufti checks it. Then another specialist checks the mufti. Difficult questions go to a group of scholars. The world’s first AI mufti had quietly been demoted to research assistant.
That was the position the UAE brought to Cairo: let AI do the work; just don’t let it sign.
Doha goes a step further. Since 2024, Mohammed Ghaly — an Azhar-trained ethicist, not an engineer — has been leading a Qatar-funded project to build IslamGPT 1.0. That complicates the neat split between theologians and builders: Ghaly is building what Faruque asked for. The model, due to run through 2027, is being designed for Arabic Islamic Studies and trained on classical and contemporary Arabic sources by an interdisciplinary team.
The interesting part is how fussy they are being about the library. Every paragraph in the training corpus is supposed to be annotated for things like its author, historical context, scholars, discipline, and genre. Queries will use semantic similarity, but the answers will be grounded in those annotated documents specifically to fight “ChatGPT-like hallucinations.” In other words: don’t just tell the model to be Islamic. Build it a library and label the shelves.
There is an awkward aspect to all this, though. The supervisors of “Islamic AI” are authoritarian governments. Egypt, the UAE, and Qatar are not exactly bastions of liberty, so “human oversight” is not really neutral. “Keep a human in the loop” sounds a little different when the human ultimately reports upward to the absolute ruler.
In 2025 the Journal of Digital Islamicate Research tested five Islamic chatbots on eighty questions of Islamic law. They misattributed Qur’anic verses. They invented hadith. They contradicted each other and themselves. Not one reached the study’s threshold for trusted educational output. The authors, like almost everyone in this story, end up asking scholars and developers to work together.
But their diagnosis goes somewhere more interesting. Religious knowledge, they argue, depends on authenticated chains of transmission (isnād). That isn’t just a fancy citation system. A chain is people vouching for people. Better retrieval can tell you where an answer came from. It can’t, by itself, tell you whether AI has any standing to pass that answer on.
On the first day of 2024, Pope Francis devoted his World Day of Peace message to AI. He was already suspicious of the noun. AI systems, he argued, reproduce fragments of human intelligence rather than the whole thing. Their outputs depend on people who design them, own them, and decide what they are for. And there would always be something computation could not swallow: “Not everything can be predicted, not everything can be calculated.”
Algorithms work by turning the world into things they can count. Francis’s worry was that humans might eventually forget what was there before. By the time he reached autonomous weapons, he warned that not only intelligence but “the human heart” risked becoming artificial. The fear was that we might start taking lessons from the machine.
A year later the Vatican built it. Antiqua et Nova, issued in January 2025, starts by refusing the industry’s favorite shortcut. AI can “perform tasks, but not the ability to think.” Human intelligence belongs to the whole person: embodied, relational, moral, spiritual, shaped by a life actually lived. The document gets almost defiantly uncomputational about what that means: “poetry and love are necessary to save our humanity.” A chatbot may imitate the products of intelligence, but it remains “a tool, not a person.” And the warning eventually becomes theological. Treat AI as some greater Other, it says, and machines start occupying the place of God. The new idol has one advantage over the old ones: this one talks back.
Then Pope Leo XIV came and went viral.
He signed Magnifica Humanitas in May 2026, exactly 135 years after Pope Leo XIII signed Rerum Novarum, which had looked at the factory and asked what industrial power was doing to workers, families, property, and political life. Leo XIV looks at AI and drags the same questions into the server room. But before he gets to the servers, he stops at the human being they are supposed to serve:
“Our relationship with life seems to be in crisis today. Everything that appears as a ‘limit’—incapacity, illness, old age, suffering, vulnerability—tends to be seen primarily as a defect to be corrected, rather than as a reality through which our humanity matures and opens itself to relationship. And yet we must remember that humanity flourishes not despite limitations, but often through them.”
Leo’s point is stranger than the usual defense of “human dignity.” The things a technological civilization learns to call defects are also places where people learn dependence, compassion, generosity, worship, and love. He is asking what happens when eliminating every human limit becomes the definition of progress. A world without weakness sounds attractive until you ask what happens to mercy.
And then Leo turns from the vulnerable body to the very powerful company. He points out that “control over platforms, infrastructure, data and computing power does not rest with States” but increasingly with major technological actors, which can set the conditions for access, visibility, and participation. This is not a side issue for the antitrust lawyers. Whoever controls the machinery increasingly shapes the world in which everyone else works, knows, chooses, and communicates. The Pope says AI must be “disarmed”: stripped of the assumption that “technical power automatically confers the right to govern.”
When he presented the encyclical ten days later, Leo said the word was deliberately strong. He meant a race that was military, economic, and cognitive at once. And sitting in the Synod Hall listening was Christopher Olah, the Anthropic co-founder whom Leo thanked personally for coming. The Church had invited a builder into the room.
In the Cambridge Companion, the chapter immediately before Chaudhary’s is by Marius Dorobantu, a Christian theologian who agrees that present chatbots can imitate the outward form of conversation without possessing the inward life that gives conversation its depth. So far, familiar. Then, like Inloes, he leaves the door unlocked: “There is nothing inherently prohibitive in Christian theology regarding the possibility of AI one day acquiring personhood, partaking in the imago Dei or becoming religious.”
That is not Faruque’s answer. Faruque says AI can never cross the line into genuine self-consciousness. Dorobantu says Christian theology need not rule out a future machine crossing some very consequential lines. Their chapters sit back to back in the same book. Neither cites the other.
Which makes the word Western start to wobble. Why is artificial intelligence filed as Western when Christian thinkers inside the West are attacking the same reduction of human beings that supposedly produced it? The answer seems to be that “the West” is doing several jobs at once. Sometimes it means a philosophical genealogy: Galileo, Descartes, physicalism, materialism. Sometimes it means Silicon Valley: American companies and American ambitions. And sometimes it means Christianity. But Descartes is not Claude, and OpenAI is not the Catholic Church.
Faruque had said it almost in passing in the fifth lecture: other people — Western people — were arriving at some of the same conclusions. He was right. The problem was that “the West” had become a junk drawer, cramming together people who were busy arguing with each other.
The student who asked this essay’s opening question was the only one willing to say the quiet part out loud. What he couldn’t see was that the line he wanted to run behind already had people on both sides of it.
They had crossed before he asked. Somewhere a person with a phone was asking a chatbot the question they were too embarrassed to ask a scholar, getting an answer, and going about their day.
A verdict is cheap. Rebuilding is expensive. It means saying what a human is in words that couldn’t have been written in medieval times. It means deciding where science ends and Islam begins instead of inheriting somebody else’s line. It means giving up the idea that a technology belongs to a civilization rather than to a species. The courtroom lets you skip all of that. Rule against AI in the morning, use it after lunch, and nothing you believe about human beings has to move.
So the wound reopened the old ones and added to them. Copernicus, Darwin and Freud were all in the room, and none of them had ever been closed.
Everything specific in this essay will be wrong soon. The apps will be replaced. Governments and companies will revise their positions. IslamGPT will ship. Or it won’t. I don’t know.
What survives is the habit. A new claim arrives, Islam gets put in the judge’s chair, and everyone waits for the ruling. Copernicus: does this threaten us? Darwin: can we accept this? Freud: where does this stop being Islamic? AI: can this thing really know? Again and again, the argument ends by drawing a border around Islam instead of fully rethinking or engaging with what’s on the other side.
The last essay in this series will have a few words for the border patrol.
