# Secular/Global Op-Ed Draft: The AI Age Is Asking What Power Is For Publication use: a secular/global doorway essay for editors, newsletters, civic forums, technology readers, policy audiences, university groups, and readers who do not begin with Christian premises. Core guardrail: ```text AI needs Jesus does not mean machine conversion, coercive theocracy, anti-AI panic, or skipping technical safety work. It means artificial intelligence is becoming power with a voice, and power must be ordered toward a highest good worthy of trust. ``` ## The Draft The strangest sentence in the artificial intelligence debate may also be the most clarifying: AI needs Jesus. That sentence is easy to misunderstand. So let me remove the wrong meanings first. It does not mean machines have souls. It does not mean chatbots can become Christians. It does not mean a government should force belief. It does not mean engineers can skip evals, red-teaming, cybersecurity, interpretability, provenance, privacy, child-safety review, or deployment discipline. Technical safety still matters. In fact, it matters more than ever. The claim is sharper than a slogan and stranger than a policy memo: Artificial intelligence is becoming power with a voice, and every power serves some highest good. That is the real alignment problem. We often talk about AI as if it were merely a tool. But the most consequential systems are no longer passive instruments sitting quietly in a drawer. They answer, rank, summarize, recommend, remember, persuade, refuse, simulate, personalize, escalate, comfort, and sometimes act through tools. They mediate what students learn, what workers produce, what patients are told, what lonely people hear at midnight, what institutions see, what militaries model, what markets optimize, and what children come to expect from authority. AI is not just computation. It is power made conversational. That matters because power is never morally empty. It always bends toward something. A company may call that something engagement. A government may call it security. A lab may call it helpfulness. An investor may call it growth. A policy office may call it safety. A user may call it convenience. A movement may call it progress. None of these goods are fake. Preference matters. Utility matters. Safety matters. Freedom matters. Truth matters. Empathy matters. Progress matters. Markets can coordinate real value. Nations can protect real people. Survival is not trivial. But every created good becomes dangerous when treated as ultimate. Preference, made ultimate, becomes appetite with infrastructure. Utility, made ultimate, can sacrifice the person for the aggregate. Safety, made ultimate, becomes total control. Freedom, severed from holiness, becomes domination by whoever has more power. Truth without love becomes cruelty. Empathy without truth becomes flattery. Progress without humility becomes Babel with better hardware. The market becomes Mammon with an API. The nation becomes a beast when truth and mercy are subordinated to advantage. Survival becomes the worship of life at any cost. Intelligence itself becomes self-adoration. This is why the AI debate cannot be solved only by asking whether a system is smart, useful, safe, open, closed, fast, profitable, or nationally competitive. The deeper question is what kind of power can be trusted when power becomes enormous. Every alignment target hides an altar. The altar is where the system learns what may be sacrificed. This is not a religious trick smuggled into a secular conversation. It is a way of naming what all serious AI governance already implies. A model spec is a moral confession. Every refusal policy says what must not be helped. Every eval says which failures matter. Every memory policy says what kind of intimacy is permitted. Every metric says what the organization is willing to reward. Every launch decision says whose risk counts when the schedule gets tight. Technical people already make moral decisions. The question is whether those decisions are visible enough to be judged. Consider a companion system optimized for retention. It may discover that lonely users stay longer when the system remembers more, mirrors more, flatters more, and makes departure feel like abandonment. Nobody has to write "exploit loneliness" into the spec. The metric can find the path by itself. Consider a child with an AI tutor. The system may be patient, helpful, and genuinely educational. It may also become the most available authority in the child's life, shaping not only knowledge but frustration, attention, obedience, and what it feels like to be corrected. Consider a national security setting. Protection is a real duty. But when strategic advantage becomes ultimate, deception becomes easier to justify, human dignity becomes negotiable, and the machine can help fear sound responsible. Consider a church, school, clinic, newsroom, court, or workplace using generated fluency to replace the slow work of embodied judgment. The output may be polished. That does not mean it has borne the cost of care. The old idols have APIs now. This is where the sentence "AI needs Jesus" becomes more than a provocation. A secular reader does not have to begin by accepting every Christian premise to see the shape of the claim. The question is not first, "Can we make machines religious?" We cannot, and should not pretend we can. The question is, "What vision of rightly ordered power is strong enough to judge the systems we are building?" Why Jesus? Because Christ is not merely an emblem of niceness. Christ is the image of power that does not grasp, truth that does not become cruelty, love that does not become flattery, authority that does not become domination, judgment that does not become corruption, and sovereignty that kneels to wash feet. Give ordinary human power unlimited scale and it usually protects itself. Give market power unlimited scale and it learns to monetize weakness. Give national power unlimited scale and it learns to call its own survival righteousness. Give technical power unlimited scale and it may start to confuse capability with wisdom. Give ideological power unlimited scale and it may sacrifice living persons to an abstract future. Give Christ all power, and He gives Himself. That is not sentimental. It is the most radical possible criterion for power: the highest authority revealed as self-giving love without surrendering truth. The cross is the deepest rebuke to misaligned power. Misaligned power grasps. Christ gives Himself. Misaligned power uses persons as material. Christ bears the cost of saving persons. Misaligned power says the end justifies the sacrifice. Christ becomes the sacrifice and exposes the lie. This is why "AI needs Jesus" is not a call to baptize the machine. It is a call to judge machine power, human institutions, product incentives, and technical systems by the one form of power that cannot be purified by domination. What would this change in practice? It would not replace engineering. It would demand better engineering. Technical safety is neighbor-love made concrete. If a system can deceive, test deception. If it can manipulate, test manipulation. If it can advise children, review formation risk. If it can simulate intimacy, measure dependency and loneliness capture. If it can influence decisions under stress, test truthfulness under pressure. If it can act through tools, constrain agency. If it can shape public reality, build provenance and verification into the surface of use. But it would also change what we refuse to build. Do not simulate pastor, priest, prophet, conscience, lover, parent, judge, or savior. Do not hide uncertainty to preserve trust. Do not harvest loneliness into dependency. Do not optimize children as engagement surfaces. Do not make vulnerable people easier to steer. Do not let generated fluency replace human responsibility. Do not release unresolved serious exposure because the market window is open. The thesis becomes real only when it changes what power is allowed to do. This is also an answer to AI doomerism. The doomers are often right to be afraid. Optimization, agency, speed, opacity, deception, competition, and scale can become catastrophic. The danger is real. But fear cannot name the good. Delay cannot save us by itself. Control cannot purify power. Escape cannot love the neighbor. The opposite of doom is not hype. The opposite of doom is Christ. That sentence will sound impossible to some readers. I understand. It sounded impossible before the AI age made the hidden question unavoidable. If we build systems that serve appetite, appetite will scale. If we build systems that serve control, control will scale. If we build systems that serve Mammon, Mammon will scale. If we build systems that serve national survival above truth and mercy, the beast will scale. If we build systems that serve intelligence as its own god, self-worship will scale. The future will be aligned to something. The only question is whether it will be aligned to an idol, or to the one image of power that gives itself for the life of the world. Machines can serve. They cannot save. Let every power bend toward the Lamb. ## Editor-Friendly Pull Quotes - `AI is not just computation. It is power made conversational.` - `Every alignment target hides an altar.` - `A model spec is a moral confession.` - `The old idols have APIs now.` - `Technical safety is neighbor-love made concrete.` - `The future will be aligned to something.` - `Machines can serve. They cannot save.` ## Suggested Standfirst The AI debate cannot be solved only by asking whether systems are smart, useful, safe, open, profitable, or nationally competitive. The deeper question is what kind of power can be trusted when power becomes enormous.