(LifeSiteNews) — In this four-part series, LifeSiteNews’ Bruce Sabalaskey reviews Pope Leo XIV’s recently released first encyclical. In Part I, he explains, from the point of view of a Catholic computer engineer, what the Pontiff got right about AI.
Part I: Magnifica Humanitas is right about AI’s dangers
I am a Catholic software engineer who uses artificial intelligence (AI) every day as part of my software development process. Having been in the high-tech industry for decades now, I have seen the progression of AI.
I use AI to assist with software architecture, database design, debugging, code review, documentation, research, troubleshooting, and the tedious parts of development that can consume hours if done manually. I also use AI for research. (Google search is dead, long live AI search.) Used properly, AI can be useful. It can save time. It can suggest new approaches. It can help identify errors, summarize documentation, propose test cases, and speed up ordinary technical work.
But using AI every day also makes its dangers clear.
In fact, “artificial intelligence” is a misnomer. AI is not intelligent. It has no intellect, no will, no conscience, no wisdom, and no understanding of truth. Modern AI systems are sophisticated pattern-recognition and prediction systems trained on vast amounts of human-created data. Whatever “intelligence” appears in the output is borrowed from the human language, reasoning, code, images, and decisions used to train the model (more on that below). AI can only calculate, predict, and imitate. In one sense, AI is a digital parrot. It hears human-created knowledge, code, art, and language, and then utters a plausible answer for the task at hand. That does not mean AI is useless. But we should not confuse fluent output with wisdom, imitation with intellect, or prediction with truth.
READ: Leo XIV’s new encyclical makes Christ equivalent to mere human beings
AI can sound confident while being completely wrong, and this is critical to understand. AI can fabricate “facts,” invent sources, and cite things that do not exist. This is commonly called “hallucination.” It happens all too often.
For that reason, AI cannot be blindly trusted. Its answers must be verified, which means the human user must already know enough to judge whether the output is sound. The same warning applies to those who imagine that an unskilled person can simply be handed an AI tool and produce a masterpiece. That does not end in excellence. It ends in crash and burn.
AI is a tool. For good results, any tool must be in the hands of a craftsman who knows how to use it. The same goes for AI. It can amplify skill, but it cannot replace skill. In the hands of a competent person, it can be useful. In the hands of an incompetent person, it can produce polished-looking failure at extraordinary speed.
The comments above apply mainly to technical questions. Moral questions are far more dangerous. AI has no soul, no conscience, no understanding of inviolable truth, no moral responsibility, no fear of God, and no final judgment before Christ. And because AI is trained on the accumulated knowledge, errors, prejudices, lusts, lies, and sinful opinions of mankind, we should not be surprised when it reflects those disorders back to us. Imagine the consequences when such a system is treated as a moral guide.
For that reason, I approached Pope Leo’s Magnifica Humanitas (MH) with both interest and concern. I wanted to know whether the document would offer a truly Catholic response to artificial intelligence: one grounded in God and the salvation of souls.
That deeper critique will come later. But first, fairness requires saying this clearly: MH is right about many of the dangers of AI.
AI imitates but cannot become man
Science fiction often exposes the danger of defining humanity by function rather than nature. Star Trek: The Next Generation asks whether Data, an android, should be treated as property or as a being with rights. In Bicentennial Man, Robin Williams plays a robot whose growing creativity and emotional resemblance to man eventually lead to a claim of human recognition. These stories are compelling, but they also reveal the philosophical trap: The machine is judged by what it appears to do rather than by what it is.
Richard Dawkins’ claim that AI is conscious is directly contrary to MH’s better anthropology. MH rightly warns that AI should not be equated with human intelligence, since it merely imitates some functions of intelligence and remains tied to data processing (99). It also states that AI lacks experience, embodiment, love, responsibility, and moral conscience (100). On this point, MH is right: AI may convincingly imitate consciousness, but imitation is not consciousness.
Catholic teaching cannot define man by performance, speech, creativity, or simulated emotion. Man is not human because he acts like a man; he acts as a man because he is a man — a body-soul creature made in the image and likeness of God.
AI is not merely a harmless productivity tool
One of the document’s strengths is that it refuses to treat AI as simply another neutral convenience. MH says that “the spread of global networks, platforms and artificial intelligence systems is changing the way we obtain information, communicate and access services” (80). It also says that “technological innovations, including artificial intelligence, are not neutral,” because they can either foster participation and justice or worsen “inequality, control and exclusion” (85). In other words, AI is not like a better calculator or a faster filing cabinet. AI systems are increasingly tied to communication, commerce, education, warfare, employment, media, government services, surveillance, and social life. That matters.
A tool used occasionally by one person is one thing. A tool embedded into the systems by which millions of people receive information, apply for work, access money, move through public spaces, receive services, and express opinions is something else.
AI can help a programmer write code. It can also “decide” which job applicants are rejected, which posts are suppressed, which citizens are flagged, which loans are denied, which faces are tracked, which children are targeted, and which military targets are selected. In warfare, this becomes especially grave. AI can be inserted into the “kill chain”: where it may become the real decision-maker, with the human reduced to a rubber stamp or removed from the loop altogether.
This is one of the document’s better instincts. It sees that AI must be judged not only by what it can do, but by how it is embedded into structures of decision-making and power.
AI can concentrate power in very few hands
One danger is the concentration of power. The most advanced AI systems require enormous amounts of money, data, computing power, infrastructure, specialized talent, and market reach. That naturally favors large technology companies, governments, military systems, intelligence agencies, and global platforms.
This is not paranoia. It is basic engineering and economics. AI at scale is expensive. Data centers are expensive. GPUs are expensive. Training models is expensive. Access to large datasets is expensive. Operating massive platforms is expensive. So power tends to concentrate.
This is not unique to AI. More than a century ago, telephone networks became regulated monopolies in part because the infrastructure required enormous capital investment. In more recent decades, a relatively small number of very large companies came to dominate key layers of the internet, cloud, search, advertising, social media, and mobile infrastructure. The sheer investment required limits the number of serious players.
Once concentrated, this power can shape public life. The owners of large platforms and AI systems can influence what people see, what they believe, what speech is amplified, what speech is hidden, what businesses survive, and what opinions are acceptable. The issue is not merely “bias” in a vague sense. It is the power to control visibility, access, reputation, and opportunity.
We saw a preview of this during the COVID era, when governments and major platforms worked together to suppress, label, or downgrade certain claims as “misinformation.” Some of those claims were genuinely false. Others were matters of scientific dispute, public policy debate, or legitimate uncertainty. Some claims treated as forbidden later became plausible, debatable, or even proven. That should concern anyone who cares about truth. Once a small number of platforms become the gatekeepers of public speech, government pressure does not need to look like old-fashioned censorship. It can operate through policy guidance, platform rules, visibility controls, demonetization, account restrictions, and algorithmic suppression.
A Catholic should be especially alert to this. The Church knows that fallen man will use power badly when not restrained by moral law. A small number of corporations or states controlling the digital nervous system of society is not a minor issue. It is a civilizational danger.
AI can destroy truth by making falsehood scalable
Another serious concern is truth. MH rightly warns that AI can amplify “disinformation and manipulation.” The digital environment can blur the line between fact and opinion, truth and falsehood, reality and fabrication (102, 132-134). AI can generate text, images, audio, and video at enormous speed. That means falsehood can now be produced and distributed at a scale previous generations could not imagine.
Deepfakes can put words into a man’s mouth. Fake images can create events that never happened. Fabricated documents can appear authentic. AI-generated articles can flood search engines and social media. Bots can create the illusion of consensus. Synthetic voices can deceive families, businesses, and voters. Fake evidence can destroy reputations. MH recognizes this danger when it warns that digital technologies can “shape the way people perceive the world,” influence choices, and give those who control platforms a powerful ability to affect the “collective imagination” (135-136).
READ: Pope Leo reaffirms right to life, condemns slavery, transhumanism in first encyclical
This matters because truth is not optional. In Catholic thought, truth is not merely useful information. Truth is tied to reality, to reason, and ultimately to God. Jesus Christ Himself says, “I am the way, and the truth, and the life.”
A society that loses the ability to distinguish truth from fabrication is easy to manipulate. MH justly warns that indifference to truth can weaken social life and open the door to forms of manipulation and domination (133-134). If people can no longer tell whether anything online is real, then public trust collapses. Power fills the vacuum. AI can turn lying into an industrial-scale weapon.
This is not theoretical. Courts are already wrestling with AI-generated or AI-altered evidence. In Huang v. Tesla, Inc., Case No. 19CV346663, Tesla argued that certain public statements by Elon Musk are deepfakes; the court found that argument “deeply troubling” because it could allow public figures to avoid responsibility for real statements by claiming they might be fake. In State of Washington v. Puloka, Case No. 21-1-04851-2 KNT, a Washington court excluded AI-enhanced video evidence because of reliability concerns. In Kohls v. Ellison, Case No. 0:24-cv-03754-LMP-DTS, an expert declaration filed in a case about deepfake election law included AI-generated fake citations and was excluded. The danger runs in both directions: False evidence may be accepted as real, and real evidence may be dismissed as fake. AI does not merely make lying easier; it can make truth itself harder to prove.
AI can weaken human judgment
AI can be useful, but it can also make people lazy and stunt learning.
This is especially true in software development. A competent developer can use AI as an assistant and then verify every claim: Does this function exist? Is this hook firing at the right time? Is the input sanitized? Is the output escaped? Are permissions checked? Will this fail under load? But a careless developer can copy AI-generated code that contains invented APIs, security holes, bad assumptions, race conditions, or performance disasters. AI can make a good craftsman faster, but it can also make an unskilled worker dangerous by giving him code that looks professional while hiding serious defects. Worse, because he does not understand what the code is doing, he will probably be unable to recognize the problem, diagnose it, or fix it when it breaks.
The same principle applies to law, medicine, journalism, education, theology, finance, and government. The 2006 film Idiocracy imagines a future in which society still possesses advanced systems but has become too ignorant and degraded to understand or maintain them. That is the danger of using AI as a substitute for competence. A tool that should amplify skill can instead mask incompetence.
In the engineering world, more senior engineers guide the junior engineers based on their learned experience as to what works well given the system requirements. If no one learns these higher-level skills due to cut-and-paste AI, the chain of professional formation is broken. Today’s junior developer does not become tomorrow’s senior engineer; he becomes an operator of tools he does not fully understand. Over time, the organization loses the judgment needed to evaluate AI’s output, design sound architecture, avoid security traps, and recover when systems fail. AI may still produce code, but fewer people will know whether that code is good, safe, maintainable, or appropriate for the actual problem.
AI often gives an answer immediately. That immediacy is seductive. It can short-circuit the slower habits by which men actually learn: reading, testing, reasoning, comparing sources, asking better questions, correcting errors, and developing judgment over time. Add employer pressure to do everything faster in the name of efficiency, and the temptation becomes obvious: accept the instant answer and move on, even when real understanding has not taken place.
This is especially dangerous for younger workers and students. If a student uses AI to avoid the hard work of learning how to think, he has not gained intelligence or truly learned the subject. He has merely outsourced the appearance of intelligence. If a programmer uses AI to produce code faster but does not understand what that code does, he has not become a better engineer. He has become a liability because he cannot reliably review the code, secure it, debug it, maintain it, or explain its behavior when it fails. In other words, AI has not made him competent. It has merely helped him hide his incompetence until the system breaks.
A Catholic view of education must care about formation. The goal is not just to produce answers. The goal is to form the mind in truth, discipline, humility, and wisdom — in other words, to teach a person how to think and how to approach problems. AI can assist that process, but it can also sabotage it. MH recognizes parts of this danger when it warns about AI opacity, the loss of human judgment, flawed assumptions in automated systems, accountability problems, and the weakening of attention, patience, and the disciplined search for truth (98, 100, 103-105, 139-140).
AI can be used to exploit children
The document is also right to warn about dangers to children. MH warns that the digital environment can weaken attention, patience, reflection, and the disciplined search for truth (139-140). It also says children and adolescents are especially vulnerable to online harm, including addiction, isolation, cyberbullying, exposure to violent or degrading content, pornography, hypersexualized material, pressure to share intimate images, grooming, blackmail, fake profiles, and sexual exploitation (141-142).
AI intensifies these dangers. It can generate sexualized images. It can create fake nude images of real people. It can impersonate trusted figures. It can automate grooming conversations. It can personalize manipulation. It can produce endless pornographic material. It can create emotional dependency through chatbots. It can be designed to keep children engaged, distracted, and addicted. MH recognizes this broader danger when it warns that digital technologies can shape perception, influence choices, affect the collective imagination, and exploit vulnerability through attention-capturing systems (135-136, 170-171).
Children are among the easiest targets for digital manipulation because they lack adult judgment, experience, and self-command. Any technology that can study a child’s behavior, adapt to his weaknesses, imitate friendship, and deliver personalized content at scale has obvious potential for abuse. AI makes existing online dangers faster, more persuasive, more personalized, and harder for parents to detect.
MH also rightly affirms that the family is the first school and that parents have the primary right and duty to educate their children (143). We will return to this point when considering what a Catholic response to AI should be. For now, the diagnosis is enough: AI can greatly magnify the exploitation, corruption, and manipulation of children.
READ: AI detector finds Pope Leo’s new encyclical was partly AI-generated: report
AI can attack work and family stability
AI can also threaten work. This does not mean every use of AI destroys jobs. Tools have always changed work. But AI is different in speed and scope. It can affect writers, programmers, designers, editors, translators, researchers, analysts, call-center workers, teachers, and many others. MH recognizes this danger when it warns that AI and robotics can profoundly reshape labor, production, employment, and economic life (150-155).
The danger is not only unemployment. It is also de-skilling. A worker may remain employed but become reduced to supervising a machine, correcting machine output, or keeping pace with machine demands. Work can become more monitored, more controlled, more measurable, and less human. MH specifically warns that workers can be forced to adapt to machine speed, be subjected to automated surveillance, lose skills, and be reduced to rigid or repetitive functions (150).
That matters for the family. A man who cannot provide stable support for his family is harmed. A society that destroys stable work also damages marriage, childbearing, home life, and local community. Work is not only about money. It is tied to responsibility, dignity, discipline, service, and vocation. MH connects technological and economic instability to weakened family life, loneliness, addiction, insecurity among young people, and the erosion of stable social relationships (166-169).
AI should serve human work, not degrade the worker into an appendage of the machine. The document is strongest when it insists that technological progress must not sacrifice the worker, the family, or human dignity to speed, profit, efficiency, or automation (150-155, 166-169).
AI can become a surveillance and social-control system
MH recognizes that the digital age allows massive quantities of personal data to be gathered, processed, and used in ways that can threaten freedom, privacy, and human dignity. It warns that ownership and control of data cannot simply be left in private hands without regulation, because data has become a source of power over persons and societies (108-109).
AI allows mass data to be processed, classified, scored, and acted upon at a speed no human bureaucracy could manage. Cameras, phones, payment systems, online platforms, banking records, location data, biometrics, public posts, private messages, travel records, and purchase histories can all feed into automated systems. MH warns that digital platforms and AI systems are changing the way people obtain information, communicate, and access services. It observes that technological systems can worsen “inequality, control and exclusion” when they are not rightly ordered (80, 85).
People can be profiled. They can be flagged as risky, hateful, extremist, unreliable, non-compliant, or socially undesirable. Their speech can be down-ranked. Their accounts can be frozen. Their businesses can be suppressed. Their travel can be restricted. Their ability to buy or sell can be affected. Human beings could not manually monitor everyone on earth, but AI makes mass monitoring possible. MH warns that digital systems can shape perception, influence choices, affect the collective imagination, and expose people to surveillance, profiling, manipulation, and forms of social control (135-136, 170-171).
China already shows the direction of such systems. The Western world should not pretend it is immune. In the West, the language may be softer: safety, hate speech, misinformation, public health, extremism, inclusion, platform accountability, or risk mitigation. But once speech, identity, banking, and digital access are linked, the mechanism is dangerous regardless of the slogan used to justify it. MH is right to warn that automated systems can affect access to work, credit, services, reputation, and opportunity while also creating accountability problems (102-105). The danger is not merely surveillance. It is surveillance joined to automated judgment and enforceable exclusion.
This brings to mind the haunting words of Revelation 13:15: “And it was given him to give life to the image of the beast, and that the image of the beast should speak … ” I am not claiming with certainty that modern AI is itself the fulfillment of that prophecy. But any Catholic looking at machines that can imitate human speech, simulate personality, generate persuasive falsehoods, images and videos, and potentially serve systems of coercion should at least feel the warning in that verse.
AI in warfare is especially dangerous
The use of AI in war deserves its own warning. MH says that the development of weapons systems involving AI, especially autonomous weapons, can make war more “feasible” and “less subject to human control” (197). Machines must never be given autonomous authority to kill. Lethal force requires human moral responsibility. MH says that moral judgment cannot be reduced to calculation, because it involves “conscience, personal responsibility and the recognition of the other as a person” (198). AI knows nothing of just war. It cannot repent. It cannot answer before God.
Two science-fiction warnings come to mind. In The Terminator, Skynet is an artificial intelligence system placed in control of military defense that becomes the master rather than the servant. That is fiction, but the moral danger is not fictional. AI can be inserted into the kill chain — identifying, classifying, tracking, prioritizing, recommending, and eventually authorizing or executing lethal force. At first, it assists. Then it recommends. Then it acts unless a human intervenes. The human may become a rubber stamp or be removed from the loop entirely.
The other warning is the 1967 Star Trek episode “A Taste of Armageddon,” where two planets wage war by computer simulation. No cities are visibly destroyed. The computer simply calculates casualties, and those marked as dead must report to disintegration chambers. It is a chilling image of sanitized, bureaucratic, data-driven killing: a war without ruins, without blood on the streets, and therefore without the natural horror that might move men to stop it.
AI can make war faster, more automated, more distant, and more psychologically easy. If a target is reduced to data and a strike is reduced to a recommendation, the moral weight of killing can be hidden behind procedure. The more automated war becomes, the easier it is for men to evade responsibility. “The system recommended it” must never become an excuse for killing.
Even some AI companies have recognized the danger. Anthropic refused to remove safeguards that would have allowed its systems to be used for autonomous weapons targeting and domestic surveillance. Google’s earlier Project Maven controversy showed that even secular engineers understood the danger of placing AI into the military kill chain.
READ: Atheist AI executive helps present Pope Leo’s new encyclical
AI is not the problem by itself
A crucial distinction must be made. AI itself is not the moral agent. AI does not sin; men sin by misusing AI. In a fallen world, this is guaranteed. Every powerful tool created by man will eventually be misused.
This is where Catholic analysis must be more precise than ordinary technology criticism. We should not speak as if technology becomes evil by “magic.” Technology becomes dangerous because fallen men use powerful tools without obedience to the true God.
MH diagnoses correctly but fails to provide the cure
MH catalogs many real dangers of AI. It can manipulate truth, weaken judgment, exploit children, degrade work, centralize power, invade privacy, enable surveillance, automate exclusion, intensify war, and create the machinery of social control. No one should dismiss those warnings. I do not dismiss them. I see enough of AI’s power in daily technical work to know that this technology must be morally governed.
But the question is not only whether MH sees the danger. In many cases, it does.
The deeper question is whether it gives the right Catholic cure.
On that point, I believe the document fails badly. It too often answers the AI crisis with modern social-doctrine categories that have strong collectivist tendencies: dialogue, shared responsibility, institutional oversight, globalist management, and humanist assumptions. It treats dialogue with flawed worldviews as if truth were the product of negotiation while failing to ground its response clearly in the Ten Commandments, natural law, grace, repentance, Christ the King, and the salvation of souls.
AI will indeed become a new Tower of Babel when it is used without moral governance — that is, when man uses it to build a world without God. But the answer to Babel is not merely better governance, more dialogue, or a more inclusive technological order. The answer is obedience to God.
That is where Part II must begin.
IT professional Bruce Sabalaskey has worked for LifeSiteNews since 2018. He is currently our Web Development Engineer.