In 2026, Leo XIV asks whether artificial intelligence will serve the human person or quietly replace human judgment with systems optimized for speed, profit, persuasion, and control.²
The encyclical is severe, but it is not frightened. It does not tell the world to smash the machine or flee the future. It says something more demanding. Artificial intelligence is not fate. It is power. And power, if it is to remain human, must answer to judgment, law, conscience, work, truth, and peace.
That is the part the sales language tries to hurry past. The public language of AI is usually written in the language of inevitability. The future is coming. Disruption is coming. Automation is coming. Synthetic media is coming. Autonomous warfare is coming. The only serious person, we are told, is the one who adapts before the next model release makes his objections obsolete.
Leo refuses that surrender. He argues that prudence, rigorous evaluation, and sometimes a slower pace do not mean opposing progress. They are part of responsible care for the human family.³
An adult knows the difference between progress and momentum. One has a purpose. The other only has speed.
The mechanism is plain enough to miss. AI converts human judgment into surfaces that look finished: the answer, the score, the summary, the risk assessment, the image, the recommendation, the target. Institutions then begin to treat those surfaces as decisions because the surface is fast, legible, and convenient. The deeper danger is not that machines may become human. It is that institutions may become less human because the machine gives them permission to stop looking.
A newsroom learns this first through evidence. A photograph can no longer be treated as a photograph because it looks like one. A quotation can no longer be trusted because it sounds like the person who supposedly said it. A document, a voice, a video, even the first account of an event can now arrive with a shadow beside it. The necessary act becomes the least glamorous one: wait, check, trace, ask who saw what, ask where the claim began.
Leo names that problem with unusual precision. AI and digital platforms, he writes, can blur the boundaries between truth and falsehood, mixing facts with opinions and exposing people to manipulated content, images, and videos. But his answer is not centralized truth control. Truthful information, he argues, does not arise from automated command. It requires verification, cross-checking of sources, and responsible argument.⁴
That is a remarkably old-fashioned sentence. It is also the most modern one in the document.
A democracy cannot live on impressions, synthetic images, managed feeds, and plausible summaries. Leo writes that democracy is not rules and procedures alone. It depends on concordance with the facts.⁵ When that concordance erodes, democratic life does not collapse all at once. It becomes tired, performative, and easier to rule through confusion.
This is where Magnifica Humanitas speaks most directly to journalism. The answer to AI-generated unreality is not to demand that readers trust a priest, a platform, a government, or a newsroom by reflex. The answer is slower and more difficult: source, evidence, date, context, motive, money, law, pressure, consequence. That work is often dull. It takes longer than a false image takes to travel. But it has one advantage the machine does not possess. It can be held responsible.
The word beneath the encyclical is not technology. It is abdication.