Magnifica Humanitas: Pope Leo XIV on AI and Adopting It Humanely
In his first encyclical, Pope Leo XIV asks for AI to be disarmed and placed at the service of human dignity. We read the core points and show how companies can translate that care into practice.

Marlos Carmo
Founder of Tolky
Updated on November 15, 2026

Share
On May 25, 2026, the Vatican publicly presented Magnifica Humanitas: on safeguarding the human person in the era of artificial intelligence, the first encyclical of Pope Leo XIV. It was signed on May 15, 2026, the 135th anniversary of Rerum Novarum, Leo XIII's encyclical that, in 1891, laid the foundations of the Church's social doctrine in the heat of the Industrial Revolution.
The name and date speak to tradition: Cardinal Robert Francis Prevost took the name Leo XIV in May 2025, and the encyclical was signed on the anniversary of Leo XIII's Rerum Novarum (the link between the two is our reading). Where Leo XIII faced the dehumanizing potential of the steam-powered factory, Leo XIV now faces the dehumanizing potential of algorithms. The parallel is our reading: AI is our "rerum novarum," the "new thing" that demands a moral response on par with the moment.
We read the encyclical from the perspective of people who deploy conversational AI in companies every day. This article distills the main points and, at the end, shows how we translate that care into product, because efficiency and humanity are not opposite poles; they are, in fact, the same requirement.
Note: the Pope's document does not address customer service. The numbers in parentheses refer to paragraphs of the official text, available on the Vatican website. The rest is our paraphrase, and the bridges to customer service practice are Tolky's reading, not the Vatican's position.

The starting point: technology is not neutral
Magnifica Humanitas is structured as an introduction, five chapters and a conclusion, totaling 245 paragraphs. The underlying premise appears right at the start: technology is not presented as a force opposed to humanity (4), nor as inherently evil (9). But the Pope is direct: in our paraphrase, technology is never neutral, because it carries the face of those who design, finance, regulate and use it (9).
That sentence undoes two lazy arguments that dominate public AI debate:
- The argument from awe ("it's just a tool, it all depends on how you use it"), which absolves the designer of the system.
- The argument from fear ("AI is going to destroy everything"), which paralyzes instead of asking for discernment.
The encyclical refuses both. Whoever designs, finances, trains and regulates an AI system imprints intent into it, even unconsciously. There is no exit through technical neutrality. There is, however, an exit through explicit responsibility.
"Disarming AI": the document's key concept
The encyclical's strongest word is just one: disarm. It carries deliberate weight:
In paragraph 110, disarming AI means removing it from the logic of armed competition, which today is not only military but also economic and cognitive (our paraphrase of the official text). Disarming does not mean rejecting technology, but preventing technical power from presenting itself as a right to govern.
The text treats AI as a technology whose effect depends on the logic in which it is embedded. Today it lives within three "armed" logics: military, economic, and cognitive, the last being the most subtle, because it operates on people's critical-thinking capacity.
Disarming AI, therefore, does not mean halting it. It means pulling it out of a mentality of competition, domination and exclusion, and placing it at the service of the common good. The point is strategic, not technophobic.
The "technocratic paradigm" and what it hides
The chapter on technology and dominance picks up a warning from Pope Francis: the technocratic paradigm (92), the mentality in which every choice is dictated by the measurement of efficiency and profit.
In our paraphrase, the text argues that the most powerful technology is not necessarily the best, and that AI can imitate a person without having moral conscience, empathy or affective and relational capacities.
There is a philosophical cut here that matters to the corporate world. Efficiency and profitability are real goods and are not being condemned. What the encyclical condemns is elevating them to the sole criterion of technological choice. When that happens, three dimensions that no language model captures are lost: moral judgment in the face of exceptions, empathy in the face of suffering, and the relational meaning of a conversation between persons.

The deeper danger: the person as a project to be optimized
Paragraph 112 contains perhaps the most uncomfortable idea in the document for any company that operates with KPIs: the temptation for human beings to think of themselves as a project to be optimized (our paraphrase).
It is not a figure of speech. It is a precise description of what happens when an operation starts treating the customer as a variable in an objective function (average handle time, NPS, cost per contact) and the employee as a resource to be tuned to the speed of the machines.
The text criticizes transhumanism and posthumanism, currents that read progress as overcoming human limits. For the encyclical, in our paraphrase, the human does not flourish despite limits but often through them (118). Tiredness, error, fragility, illness, old age: none of these are defects to be eliminated. They are part of what makes someone worthy of unconditional respect.
The practical consequence for companies: an AI system may relieve what tires, but it cannot treat tiredness as a flaw in the human.
The dignity of work in the "fourth industrial revolution"
In the fourth chapter, the Pope addresses what is perhaps the most sensitive topic for B2B: what AI does to work.
The passage (150) is blunt, in our paraphrase: the "new ways" of working are not necessarily better, and AI, while promising productivity, often forces workers to adapt to the speed and demands of machines instead of machines being designed to support those who work.
The inversion matters. Automation should be designed to support the human, not the other way around. When a service representative spends an entire shift hitting average-handle-time targets that only make sense at bot scale, work has stopped being human: it has become an interface for the system.
The encyclical does not condemn automation. It condemns automation that cuts costs by eliminating jobs without offering a path. Technology can, in fact, free human beings from painful or repetitive tasks, provided that liberation translates into more qualified, better-paid and more meaningful work, not into unemployment.
Truth, communication and the risk of the "architecture of visibility"
The encyclical calls for an ecology of communication grounded in truth. It is concerned with:
- Transparency in how content is selected
- Personal data protection
- Serious journalism, built on argumentation and verification
- The proper and critical use of digital tools
The text speaks of an "architecture of visibility" (171): the design of digital platforms that amplifies only what is visible and shapes opinions (our paraphrase). The warning applies to any company that adopts generative AI in customer communication: what the system chooses to show (and in what order) is not neutral, it is an implicit editorial decision.
For the Pope, the massive collection of data and the use of algorithmic systems can create "a new power" (171), capable of discriminating against the most vulnerable. The defense is structural: education for critical use, independent oversight, clear legal frameworks.

What Magnifica Humanitas asks of those who build AI
Anyone reading the encyclical looking for a list of "prohibitions" walks away empty-handed. The text is, in fact, a plea for intentionality. The asks that, in our reading, the text makes of developers and companies that use AI:
| Axis | Encyclical's request | Practical implication |
|---|---|---|
| Accountability | Clarity about who answers at each stage of the AI lifecycle | Logs, audit trails, traceability of automated decisions |
| Ethical code | Shared social-justice standards, not decided by a few (107) | Public usage policy, committees with real plurality |
| Oversight | Independent oversight of systems | Active human curation, not just "sample reviews" |
| Education | Training for proper and critical use of tools | Customer and employee training in healthy use |
| Sustainability | Do not overlook AI's environmental impact (energy, water) | Assess the ecological cost of each model deploy |
| Work | Person-centered systems, not performance-centered | Metrics that include relationship quality, not only volume |
One synthesis appears in paragraph 107, which warns about the risk of AI's morality being decided by a few. AI ethics is not a problem to be solved by big tech alone: it is a social problem.
How we read this document at Tolky
We are a B2B conversational AI platform, and our daily work is to make customer service, sales and relationship management more efficient. That is why reading Magnifica Humanitas is, for us, not an academic exercise: it is product raw material. This section is Tolky's reading, not the content of the encyclical.
There is a comfortable cliché in the market: saying "technology replaces people" as if that were evidence of modernity. We reject that cliché. We believe well-built AI does not replace the human: it frees the human for what humans do best: exceptions, empathy, judgment, creativity, negotiation, care.
Picture a hypothetical scenario: when an AI resolves the twelfth duplicate-invoice request of the day, it is not "eliminating a job." It is freeing that agent for the customer who is crying on the phone because his father died and the insurance denied the procedure. That customer needs a person. For more on this, see our article on how to implement AI in customer service without losing the human touch, and on AI governance in customer service.
The principles we apply
Concretely, we translate the encyclical's care into five product decisions:
1. Human handoff with context, not as a system failure
In our view, transferring a conversation to a human agent is not "the AI didn't make it." It is the AI doing the right thing. The platform can be configured with escalation rules (for example frustration, complexity, relationship value, a policy exception) to route the conversation before the customer has to ask. The human agent receives the context: history, intent and actions already taken.
2. The machine adapts to the person, not the other way around
The idea in paragraph 150, that machines should support those who work, guides our operator-panel design. Rhythm, priorities, queues and AI suggestions exist to reduce the agent's cognitive load, not to push them to the limit.
3. Measurable efficiency, but with human-quality metrics
We optimize first response time, deflection and cost per contact because those metrics matter to operations. But we pair them with relationship-quality metrics that we suggest tracking: CSAT segmented by interaction type, repeat-contact rate within 24h, post-handoff NPS, sentiment across the conversation. Efficiency is only good when it comes with relational quality.
4. Governance and audit as default
The text calls for independent oversight. In our product, traceability translates into immutable history, audit logs on all plans and end-to-end encryption, with data handling under Brazil's LGPD; SSO is available on the Enterprise plan (details at /en/precos). Anyone using the platform can know who said what and when, and review it.
5. Limits where the human always decides
Not everything should be automated, even when it is technically possible. That is why we advocate designing "human-mandatory" zones: types of request, customer profiles, value ranges, sensitive content, set up through rules and flows. The AI prepares, but does not close. That is Tolky's reading, not the Pope's: to disarm AI is to decide, in each operation, where technical power must not replace human responsibility.
The question the encyclical hands back to every company
In paragraph 129, the text says that the true alternative is not between enthusiasm or fear, but between two ways of building (our paraphrase: a progress that serves people and peoples, or one that subjects them to the logic of power).
It is a demand. Anyone deploying AI in customer service, sales, collections, HR, or anywhere in the operation has to answer: am I designing this system to serve the people who will touch it, or am I designing it to extract efficiency from them?
Magnifica Humanitas is a call to intentionality. It is not a manifesto against productivity. It is a reminder that productivity built on top of people who are treated as projects to be optimized destroys the hardest capital to rebuild: trust.
Paragraph 126 points the same way: technology can relieve suffering and open possibilities, but must not deny the essence of the human, tied to relationship and love (our paraphrase).
In one sentence: the best AI is the one that amplifies what is best in people, not the one that tries to take their place. That is the product we are building.
Want to structure a conversational AI operation that scales without dehumanizing? Talk to our team or see the plans: together we map the right point between operational efficiency and respect for the person on the other side of the conversation.
Share
Tags
- magnifica humanitas
- pope leo xiv
- encyclical
- ai ethics
- humanization
- responsible ai
- dignity of work
- #conversational ai
About the author

Marlos Carmo
Founder of Tolky
Marlos Carmo is an AI entrepreneur and founder of Tolky, the conversational-era infrastructure and AI CRM that unifies intelligent service, multi-channel support (such as WhatsApp and voice), live CRM, and operational intelligence in a single ecosystem. He is a finalist for the SXSW Innovation Awards and a member of Francesco's Economy, a global network of young entrepreneurs focused on innovation and social impact. He works connecting Artificial Intelligence and digital transformation in projects for large organizations.
Next step
Want to bring this to your operation?
Tati, Tolky's AI assistant, understands your scenario in a few questions, suggests the best path and connects you with someone from the team, with full context.


