For nearly two thousand years, every age has tried to recognize the Antichrist in the face of its greatest fear. Our age may be doing the same with artificial intelligence, although this time the comparison raises an uncomfortable question that cannot be dismissed quite so easily: what if the danger is not that a machine becomes a supernatural being, but that human beings gradually create a system to which they surrender the authority once reserved for gods, kings and prophets?
There is something strangely familiar about the anxiety surrounding artificial intelligence today. The technology is new, astonishingly new in the form most people now encounter it, yet the fear it provokes belongs to an ancient part of the human imagination. Whenever societies have entered periods in which the familiar world seemed to be slipping away, people have searched old prophecies for signs that history was approaching some predetermined end. Wars, epidemics, political upheavals, economic collapse, nuclear weapons and cultural revolutions have all, at one time or another, been interpreted as evidence that the final chapter had begun.
Now artificial intelligence has entered the prophecy.
A recent BBC investigation brought together pastors, theologians and researchers who have begun asking publicly whether AI could play some role in Christian scenarios of the “end times.” American pastor Philip Anthony Mitchell has argued that artificial intelligence will have a major role in the events preceding the end of the world, while pastors Greg Laurie and Mike Signorelli have entertained similar possibilities, Signorelli explicitly connecting modern technology with the path toward the Antichrist. None of them denies that AI may be useful, and none can demonstrate that a biblical prophecy is actually describing modern computing, but their concerns have found a receptive audience because they touch something deeper than theology alone.
The idea sounds sensational when reduced to the headline “AI is the Antichrist.” It becomes considerably more interesting once that headline is discarded.
The Book of Revelation does not mention computers, algorithms, artificial intelligence, biometric identification or digital currencies. More surprisingly, the word Antichrist itself does not appear in Revelation at all. The term occurs in the First and Second Epistles of John, where it is connected principally with opposition to Christ, false teaching and deception. Revelation presents other symbolic figures instead: the dragon, the beasts, Babylon and, later, the false prophet. Christian tradition gradually brought several of these strands together, which is one reason popular culture now tends to treat “the Beast” and “the Antichrist” almost as though they were interchangeable characters.
That distinction is more than a theological technicality, because once it is understood, the entire question begins to change. Perhaps it makes little sense to search Revelation for a hidden description of a computer that John of Patmos could not possibly have named. Perhaps the more intriguing approach is to examine what Revelation feared: concentrated power, deceptive signs, manufactured worship, economic dependence and a system in which loyalty becomes inseparable from ordinary life.
Those themes are considerably harder to dismiss.
The ancient beast was born in a very different world
Revelation emerged from the world of the Roman Empire, and its extraordinary imagery makes far more sense when that world is remembered. Yale’s New Testament scholarship describes the book as deeply hostile to Roman imperial power, with “Babylon” functioning as a symbolic name for Rome and the beasts representing forms of oppressive empire. The famous number 666 has long been associated by many scholars with Nero Caesar through gematria, an ancient practice in which letters were assigned numerical values. Even the alternative reading of 616 found in some early manuscripts fits a different spelling of Nero’s name, one reason the Nero interpretation has remained so influential.
For the first Christians hearing Revelation, Rome was not an abstract political theory. It was the civilization surrounding them, a vast system of military authority, commerce, taxation, public ritual and imperial symbolism. Coins carried the emperor’s image; cities displayed imperial power everywhere; religion, economy and political allegiance could overlap in ways that modern secular societies often find difficult to imagine.
One Yale Bible Study commentary even interprets Revelation’s language about buying and selling in connection with Roman coinage bearing imperial images, although, like much in Revelation scholarship, the details remain debated. What matters for our present discussion is that the text originally spoke to people living inside a system of power they could neither easily escape nor openly defeat.
That historical reading does not prevent Christians from believing Revelation may contain a further prophetic dimension. It simply warns us against assuming that every new technology was secretly encoded in its pages.
Yet history has rarely stopped people from trying.
Napoleon was identified with the Beast. During other periods, religious interpreters found the Antichrist in popes, emperors, political leaders, revolutionary movements and ideological enemies. Jeffrey C. Pugh, interviewed by the BBC, recalled that his own grandmother once believed Franklin D. Roosevelt was the Antichrist. The New Deal, arriving amid the terror and dislocation of the Great Depression, appeared revolutionary enough to some Americans that biblical language seemed capable of explaining what ordinary politics could not.
During the Cold War, the Soviet Union and communism occupied similar territory in some apocalyptic interpretations. After September 11, different fears appeared. The candidate changes because the world changes.
Artificial intelligence could therefore be only the latest object onto which humanity has projected its oldest anxieties.
Except there is a difference that makes this particular candidate unusual.
We did not invite Napoleon into our bedrooms to answer questions before we went to sleep. We did not ask the Soviet Union to compose our letters, explain our symptoms, recommend our purchases, teach our children, organize our thoughts and help us decide what was true.
AI is not standing outside the gates.
We are carrying it inside.
The strange attraction of the end of the world
The belief that history is approaching its final crisis remains much more widespread than many people realize. A large Pew Research Center survey conducted in the United States in 2022 found that 39 percent of American adults believed humanity was living in the “end times.” Among Christians the figure rose to 47 percent, while 63 percent of evangelical Protestants and 76 percent of Protestants in historically Black churches expressed that belief. The survey involved more than 10,000 American adults, so these were not merely the opinions of a tiny apocalyptic subculture.
There is also a long psychological and historical pattern behind such expectations. Noreen Herzfeld, a theologian who has spent years studying the relationship between religion and artificial intelligence, observes that millenarian movements tend to flourish during periods of social stress, when people feel that events are accelerating beyond their control. In a 2024 chapter for The Cambridge Companion to Religion and Artificial Intelligence, she makes an especially interesting point: modern society projects both salvation and catastrophe onto AI. Some imagine that advanced intelligence will solve disease, scarcity and perhaps even mortality, while others imagine a superintelligence escaping human control and destroying the civilization that created it.
In that sense, artificial intelligence has already acquired something resembling an eschatology of its own.
The religious believer may imagine the return of Christ, judgment and the end of the present age, while technological futurists speak about artificial general intelligence, superintelligence, radical life extension or a transformation after which human civilization may never again resemble what came before. The vocabulary is different, yet both stories place humanity on the threshold of an event so large that history will be divided into a before and an after.
It would be too simple to say that technological futurism is merely religion disguised as engineering, but the resemblance is difficult to ignore. Humans seem remarkably attracted to stories in which ordinary history suddenly opens into something greater, whether that something is salvation or destruction.
Perhaps what frightens us most about AI is not that we know what it will become, but that we do not.
Revelation was concerned with deception before it was concerned with technology
If there is one element of ancient apocalyptic imagery that feels unexpectedly contemporary, it is not the number 666, nor the famous mark, nor even the terrifying beast rising from the sea. It is deception.
The second beast of Revelation performs convincing signs and persuades people to give their allegiance to the first. Later Christian interpretation associates this figure with the false prophet. Whatever theological interpretation one accepts, the narrative understands that power does not always rule through violence alone; sometimes it rules by controlling what people believe they have seen.
For most of human history, creating convincing false realities required enormous resources. A forged document could deceive a few people, propaganda could manipulate millions and cinema could create illusions, but each medium retained certain boundaries. A photograph was generally connected to an event that had occurred in front of a camera. A recognizable voice usually belonged to the person speaking. A moving image of a public figure ordinarily meant that the figure had, at some point, stood before a lens.
Those assumptions are disappearing.
Generative systems can create faces belonging to people who never existed, reproduce voices, manufacture photographs and generate increasingly convincing video. They can write in the style of a particular person and sustain conversations that feel spontaneous and personal. The barrier separating evidence from fabrication has not vanished completely, but it has weakened enough that the old sentence “I saw it with my own eyes” no longer carries the certainty it once did.
More troublingly, deception does not always require an AI system to invent something deliberately. The U.S. National Institute of Standards and Technology warns that generative systems can produce “confabulations,” confidently presenting incorrect or fabricated information in ways that users may mistake for reliable knowledge. NIST specifically notes that apparently plausible explanations or invented citations can encourage people to place inappropriate trust in erroneous output.
This changes something fundamental in the relationship between humans and information.
The machine does not need to know that it is deceiving us. It does not require hatred, ambition, spiritual corruption or consciousness of any kind. A system designed to predict plausible language can produce false information without possessing any concept of truth, while a human reader can nevertheless experience its answer as authoritative.
That distinction becomes important because we instinctively associate deception with a deceiver. Someone lies because someone intends to lie. Artificial systems introduce a stranger possibility: deception can emerge without a mind behind the individual sentence intending anything at all.
The effect on the person receiving the information may be the same.
An oracle that speaks in our own language
Human beings have always searched for something that knows more than they do. Ancient people travelled to temples and oracles, studied omens and consulted priests because uncertainty is one of the conditions we tolerate least comfortably. We want to know what will happen to us, whether our choices are correct, whether danger is approaching and whether a hidden order exists behind events that otherwise appear random.
Modern societies may believe they have left the oracle behind, but perhaps we have merely changed its architecture.
Instead of climbing the road to Delphi, we open an application.
The comparison should not be pushed too far, because contemporary AI is built from mathematics, software, data and hardware rather than divine revelation. Yet psychologically the encounter can become surprisingly similar. A question is offered to an entity whose internal processes most users do not understand, and an answer returns almost instantly from a body of information no individual human could possibly hold in memory.
The voice is calm. It rarely hesitates in the manner a frightened or uncertain person hesitates. It can explain complex subjects in language tailored precisely to the person asking.
The experience creates an illusion of presence that earlier technologies rarely possessed.
A search engine gave us pages to investigate. A book gave us the thoughts of an identifiable author. A calculator returned numbers. Conversational AI does something subtly different because it answers us, adjusts to us and continues the exchange. Even when we know intellectually that the system is a machine, ordinary social instincts begin to operate.
And those systems can persuade.
A 2025 study published in Nature Human Behaviour examined debates between people and GPT-4. When the AI had access to limited personal information about the human participant and the contest produced a difference in persuasion, the personalized AI was more persuasive substantially more often than its human counterpart. Other experiments published in Nature Communications have also found that messages generated by large language models can shift attitudes on contested issues and can do so at enormous scale and comparatively low cost.
None of this proves a supernatural prophecy.
It proves something more immediate: a machine does not need consciousness in order to influence belief.
The moment when assistance becomes authority
Imagine that artificial intelligence continues improving without any dramatic “awakening,” without consciousness suddenly appearing inside the machine and without some cinematic moment when humanity realizes that its creation has become alive. The systems simply become more useful, more accurate and more deeply integrated into ordinary institutions.
A physician consults an AI system because it detects patterns in medical data that a human might overlook. A financial institution uses another because fraud can be identified more efficiently. Courts experiment with AI assistance because thousands of pages of legal material can be reviewed in seconds. Schools adopt personalized systems because each student can receive individual attention. Companies rely on algorithms to screen applications, governments automate administrative decisions, intelligence services analyze threats and military organizations use machine assistance because modern warfare moves faster than human deliberation.
There need be nothing sinister in any individual step.
That is precisely what makes the process worth examining.
If a system is correct more frequently than the human professional using it, ignoring its recommendation may gradually become difficult to defend. Eventually a doctor might need to explain why she rejected the machine rather than why she accepted it. An employee may discover that questioning an automated recommendation is interpreted as inefficiency. A citizen may be told that no person can reverse a decision because the relevant procedure has become too complex for any individual to understand completely.
Nobody would need to declare that the machine was now in charge.
Authority could accumulate through convenience.
This possibility is one reason UNESCO’s global Recommendation on the Ethics of Artificial Intelligence places so much emphasis on human responsibility and oversight. Its framework explicitly states that AI systems should not displace ultimate human responsibility and accountability, precisely because a technologically produced decision can affect real people even when responsibility becomes difficult to locate.
The warning is almost mundane compared with Revelation, yet its implications are profound.
A society can surrender authority without ever holding a ceremony of surrender.
The mark, the marketplace and the seduction of modern parallels
No passage in Revelation has generated more technological speculation than the description of the mark without which people cannot buy or sell. Over the decades, interpretations have attached themselves to barcodes, credit cards, implanted microchips, biometric systems, digital identification and, more recently, digital currencies and AI-controlled financial infrastructure.
The pattern itself should make us cautious. When almost every generation discovers that its newest technology appears to fit the same prophecy, the prophecy may be functioning partly as a mirror.
Yet there is still something worth examining beneath all the speculation.
For most of history, human existence was only loosely connected to centralized technological systems. A person could possess money without an account, communicate without an electronic identity, travel without leaving a permanent digital trail and purchase necessities without being authenticated by a network.
That world is disappearing, not because of a secret apocalyptic plan but because digital systems are extraordinarily useful.
Banking becomes safer when fraud can be detected. Governments function faster when identities can be verified. Borders become more manageable through electronic documents. Businesses reduce costs through automation. Consumers prefer immediate transactions over complicated ones.
Every individual reason is understandable.
Together, however, they produce something historically new: a society in which participation in ordinary life increasingly depends upon technological systems that most citizens cannot inspect and no single individual controls.
The resemblance to Revelation does not establish a prophetic connection, but it helps explain why that connection feels so powerful to some people. Ancient texts described economic participation becoming entangled with allegiance to an oppressive power; modern societies are building increasingly integrated infrastructures through which identity, finance and access can technically be managed on an unprecedented scale.
The supernatural interpretation remains a matter of faith.
The infrastructure is real.
Perhaps the machine never needs to become evil
Science fiction has taught us to wait for the wrong moment.
We imagine a laboratory where an AI suddenly becomes conscious, recognizes humanity as inferior and begins plotting its escape. The danger has a beginning, a personality and an intention, which paradoxically makes it comforting because there would be an identifiable enemy.
History usually produces something messier.
Institutions can harm people without anyone waking in the morning intending to become evil. Financial systems can destroy lives without feeling greed. Bureaucracies can become cruel without any individual bureaucrat deciding to create cruelty. Propaganda can reshape an entire population even though paper and ink possess no consciousness.
Why, then, would artificial intelligence require self-awareness before becoming dangerous?
A system can recommend the wrong medical decision without wanting anyone to suffer. It can reproduce discrimination without understanding prejudice, amplify false information without believing it and persuade an individual without knowing that an individual exists in any human sense of the word.
Effects do not require intention.
This may be one of the most important distinctions lost in discussions about AI and the Antichrist. Those who imagine a demonic superintelligence often assume that the machine would somehow need to choose evil. The more realistic danger may be a network of systems doing exactly what they were designed or incentivized to do while producing consequences no one intended and no single person can easily stop.
In theological language, evil is often personalized.
In technological systems, catastrophe can be procedural.
The idol has never needed to be alive
Christian scripture repeatedly returns to the problem of idolatry, and the popular imagination tends to picture ancient people kneeling foolishly before statues of wood or stone. Yet the deeper idea is more sophisticated: human beings create something with their own hands and then attribute to their creation an authority greater than their own.
Seen from that perspective, the material from which the idol is constructed is almost irrelevant.
Stone will do.
Gold will do.
Silicon would do just as well.
Noreen Herzfeld’s work on religion and artificial intelligence has repeatedly returned to the question of image, creation and what happens when humans build intelligent-seeming entities in their own likeness. Her recent exploration of AI as a millenarian project is particularly revealing because she rejects both extremes: the machine as automatic savior and the machine as inevitable apocalypse. We are projecting ancient hopes and fears onto a technology whose future will still depend heavily upon choices made by humans.
This may be where the theological argument becomes more interesting than the claim that AI itself is Satanic.
An idol does not need supernatural power for people to worship it. It needs only to be treated as though it possesses an authority that should not be questioned.
Modern AI systems are demonstrably fallible, yet their linguistic fluency can make uncertainty difficult to perceive. They can sound knowledgeable where their knowledge is incomplete and confident where confidence is unjustified. As their capabilities increase, the temptation to confuse intelligence with wisdom may become stronger.
Those are not the same thing.
A system may process more information than any living person and still possess no lived experience, moral responsibility, mortality, love, grief, conscience or understanding of consequence in the human sense.
Yet if its answer is better often enough, people may stop caring about the distinction.
That is the moment worth watching.
Perhaps we have always misunderstood what obedience would look like
When people imagine the Antichrist, they often imagine a charismatic tyrant demanding worship beneath enormous banners while terrified populations are forced to obey. The image is theatrical because apocalyptic literature itself is theatrical.
Modern power often behaves differently.
The most efficient systems rarely need to threaten everyone who uses them. They become indispensable.
We did not have to be forced into smartphones. We wanted them. Social networks did not invade our homes; we created accounts. Recommendation algorithms did not demand that we surrender hours of attention; they learned how to keep us interested.
Convenience succeeded where coercion would have failed.
Artificial intelligence may follow the same path. We will use it because it saves time, then because everyone else uses it, then because institutions expect it and eventually because functioning without it becomes difficult.
At no particular point will the transformation necessarily feel dramatic.
That is what makes it mysterious.
Human beings tend to imagine historical turning points as moments that everyone recognizes while they are happening, yet many of the deepest transformations become visible only afterward. People living through the first decades of industrialization did not wake each morning announcing that they were inside the Industrial Revolution. The old world was simply disappearing around them, one machine, one factory and one migration at a time.
Perhaps something comparable has already begun.
When secular science and religious prophecy begin asking the same question
There is an irony hidden inside all of this. Religious believers are not the only people imagining catastrophic futures involving AI.
Researchers, engineers, philosophers and governments use an entirely different vocabulary, but they are also discussing scenarios in which advanced systems could create severe harm, manipulate populations, undermine institutions or behave in ways humans struggle to control. NIST develops risk frameworks, UNESCO insists upon human oversight, and researchers experimentally study whether AI can persuade human beings at scale. These are not churches interpreting Revelation; they are institutions confronting technical and social risks.
The religious language speaks of deception, false authority and allegiance.
The technical language speaks of misinformation, persuasion, alignment, accountability and loss of human control.
They are not describing the same thing, and it would be misleading to pretend they are.
Yet they sometimes appear to be standing on opposite sides of a strange bridge, staring toward the same dark water below.
Both are asking what happens when human beings create a power they do not fully understand.
Both are asking whether control can be surrendered before those surrendering it recognize what they have done.
Both are asking, in very different languages, whether something created to serve us might eventually reshape the conditions under which we live.
The question Revelation leaves behind
There is no credible evidence that artificial intelligence is the Antichrist, nor does Revelation contain a hidden technical blueprint describing computers two thousand years before their invention. Historically, much of the book’s imagery makes sense within resistance to Roman imperial power, while the term “Antichrist” itself belongs to other New Testament writings. Treating every new technology as the fulfilment of an ancient prophecy usually reveals more about the fears of the interpreter than about the prophecy itself.
But perhaps dismissing the entire comparison would be too easy as well.
Revelation survived for almost two millennia not because every generation correctly identified the Beast, but because the book speaks in images that touch persistent human problems. It is fascinated by power that wishes to become absolute, by appearances that deceive, by economic systems that can reward obedience and punish exclusion, and by the strange moment when political authority stops demanding only compliance and begins demanding something resembling worship.
Artificial intelligence did not create any of those dangers.
Human beings did.
What AI changes is their possible scale.
For the first time, a single technological system can potentially speak individually to millions of people, adapting its language to each of them while drawing upon more information than any human adviser could possess. Synthetic media can manufacture events that never happened. Digital infrastructures can connect identity, communication and economic life. Algorithms can make consequential judgments without the people affected understanding how those judgments were reached.
None of this proves the apocalypse.
Perhaps the opposite is true, and AI will eventually become an ordinary technology whose extraordinary early years seem almost quaint to future generations. Perhaps the fears surrounding it will join the long archive of moments when humanity mistook transformation for the end of the world.
Yet there is another possibility that requires no demons, no supernatural machines and no secret fulfilment of prophecy.
We may simply become so accustomed to asking machines what to read, what to believe, whom to trust, how to work, what to choose and eventually how to understand the world that the boundary between assistance and authority fades without anyone noticing exactly when it disappeared.
If that happens, the question will no longer be whether artificial intelligence has become conscious, whether it possesses a soul or whether some ancient prophet foresaw silicon chips hidden inside a vision of beasts and dragons.
The question will be far older, and much more human.
It will be whether we still remember that the thing speaking to us was created by us, whether we still possess the courage to doubt an answer delivered with perfect confidence, and whether we recognize the moment when a tool that once waited for our commands begins quietly determining the world in which those commands are possible.
Perhaps the most disturbing scenario was never that humanity would build a machine and discover Satan hiding inside it.
Perhaps it was that we would build a machine ourselves, teach it to speak with the accumulated voices of humanity, place it at the center of our institutions, consult it because it was useful and eventually trust it because everyone else did, until one day we could no longer tell whether we were still using the oracle or whether the oracle had begun, almost imperceptibly, to define reality for us.
Sources & Further Reading
BBC News — Lebo Diseko, AI and the Antichrist: Why Some People Believe Artificial Intelligence Is Deceiving Us Into Worshipping Satan, 26 September 2026. BBC News article
Pew Research Center, In U.S., 39% of adults believe humanity is “living in the end times”, December 2022. Survey of 10,156 U.S. adults; 47% of Christians and 63% of evangelical Protestants surveyed said they believed humanity was living in the end times.
Noreen Herzfeld — Cambridge University Press, The Eschatological Future of Artificial Intelligence: Saviour or Apocalypse?, in The Cambridge Companion to Religion and Artificial Intelligence, 2024. Herzfeld examines AI as a modern millenarian project onto which societies project both hopes of salvation and fears of catastrophe.
Open Yale Courses — Yale University, Apocalyptic and Resistance, New Testament History and Literature, Lecture 23. The lecture discusses Revelation in the context of Roman imperial power and explains why many scholars associate 666 — and the manuscript variant 616 — with Nero.
Encyclopedia of the Bible — Bible Gateway, Antichrist. The term “Antichrist” occurs specifically in the Johannine epistles rather than in the Book of Revelation, an important distinction from later popular interpretations that merge several biblical figures.
National Institute of Standards and Technology (NIST), Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile, 2024, updated by NIST in 2026. The report discusses risks including AI “confabulation,” in which generative systems can confidently present erroneous or fabricated information.
UNESCO, Recommendation on the Ethics of Artificial Intelligence. UNESCO emphasizes transparency, accountability and, importantly for our article, the principle that AI systems should not displace ultimate human responsibility and human oversight.
Salvi, Ribeiro, Gallotti & West — Nature Human Behaviour, On the conversational persuasiveness of GPT-4, 2025. In controlled debates, personalized GPT-4 was more persuasive than human opponents in the study’s key comparison, illustrating why AI-driven persuasion deserves serious attention even without assuming consciousness or malicious intent.
