Magnifica Humanitas – Something happened this week that most technology publications missed entirely, or buried beneath the more immediately lucrative conversation about quarterly earnings and model benchmarks. On May 25, 2026, Pope Leo XIV released Magnifica Humanitas — his first encyclical, formally titled On Safeguarding the Human Person in the Time of Artificial Intelligence — and it landed as one of the most substantive moral documents anyone in a position of global institutional authority has produced on the subject. Not because it comes from the Vatican. Because it asks the right question, names the right problem, and does so without the kind of techno-optimist softening that has characterised most mainstream commentary on AI governance.
The encyclical calls on society and AI developers to implement “shared standards of social justice” in order for artificial intelligence to respect human dignity and serve the common good. It makes clear that AI is not a morally neutral tool — that it matters not only how AI is used, but how it is designed. The document also warns explicitly that “a more moral AI is not enough if that morality is determined by a few … as with every major technological shift, AI tends to amplify the power of those who already possess economic resources, expertise and access to data.” 1
That last observation is not a theological abstraction. It is a precise description of what is happening in the insurance industry right now — and in healthcare, financial services, and every sector where algorithmic systems are making consequential decisions about real people’s lives. The fact that one of the oldest moral authorities on the planet has chosen AI as the subject of its most significant formal teaching document in 2026 is not incidental. It is a signal that the ethical dimensions of this technology have outgrown the capacity of technical and regulatory frameworks to contain them alone.
Magnifica Humanitas and the Theology of the Algorithm
It helps to understand what Magnifica Humanitas actually is before situating it in a practical context. Pope Leo XIV signed the encyclical on May 15, 2026 — the 135th anniversary of Rerum Novarum, the landmark social encyclical of Pope Leo XIII that addressed the industrial revolution’s impact on workers and the poor. The choice of date was deliberate. Magnifica Humanitas positions AI as the defining social challenge of the present moment in the same way that industrialisation was the defining social challenge of 1891. The parallel is exact and uncomfortable: industrial capitalism created immense wealth, eliminated enormous amounts of physical labour, and simultaneously produced conditions of exploitation, poverty, and systemic injustice that required a century of regulation, litigation, and social movement to partially correct. AI is doing something structurally similar, at greater speed. 2
The encyclical’s 245 paragraphs are broken into five chapters. The first two establish the development of the Church’s social doctrine from Leo XIII to the present and lay out its foundational principles. Chapter three introduces what Leo XIV calls “the technocratic paradigm” of artificial intelligence and the imbalance of digital power. Chapter four turns to the importance of safeguarding truth, democracy, work, and human relationships in the AI era. The document acknowledges familiar concerns including job insecurity, manipulation of information, privacy violations, ideological bias, autonomous weapons, and a futuristic vision of an “enhanced human being” — but identifies a deeper danger: that human beings may begin to see themselves and others as purely processable data. 3
The encyclical’s underlying premise is that technology is not “a force antagonistic to humanity,” nor is it “inherently evil.” However, “technology is never neutral, because it takes on the characteristics of those who devise, finance, regulate, and use it.” That sentence is the crux of everything. The algorithm doesn’t care about you. The people who built it, funded it, and chose which optimisation objective to train it on — those decisions carry moral weight. And in the insurance industry, those decisions are being made with commercial objectives that do not naturally align with the dignity of the person being assessed. 4
Where the Encyclical Meets the Insurance Machine
Read Magnifica Humanitas alongside what has been documented about AI-driven insurance pricing and the alignment becomes almost uncomfortable in its precision. The encyclical warns about AI amplifying the power of those who already possess economic resources and data. Our analysis of AI bias in insurance and whether algorithms can discriminate against consumers documents exactly that dynamic: models trained on data encoding structural inequality systematically overcharge communities of colour, using proxy variables like ZIP codes and credit scores to achieve racially disparate outcomes without ever invoking race directly. The algorithm amplifies existing disadvantage. The person at the end of the underwriting decision receives a premium that encodes generations of discrimination — and has almost no practical recourse.
Pope Leo XIV’s message to the Pontifical Academy for Life in November 2025 is directly relevant here. Speaking to medical professionals on AI and human dignity, he warned that “we not only run the risk of losing sight of the faces of the people around us, but of forgetting how to recognize and cherish all that is truly human.” He insisted that healthcare professionals “have the vocation and responsibility to be guardians and servants of human life, especially in its most vulnerable stages” — and that the same obligation extends to those responsible for using AI in their fields. Insurance, which determines access to healthcare, housing, and financial security, is exactly this kind of field. When an algorithm denies a health claim or prices a family out of home coverage, the harm is not statistical. It has a face. 5
The Vatican’s earlier document Antiqua et Nova, issued under Pope Francis in January 2025, established the groundwork Magnifica Humanitas builds on. That document explicitly insists that “artificial intelligence should be used only as a tool to complement human intelligence rather than replace its richness.” It warns against the risk of humanity becoming “enslaved to its own work” and argues that the term ‘intelligence,’ applied to AI, “can prove misleading” — because what AI does is pattern recognition at scale, not the creative, moral, and relational intelligence that defines human judgment. The distinction matters enormously in insurance underwriting, where the question being asked — is this person worth protecting? — is fundamentally a moral and social one, not a mathematical one. When we outsource that question to a model, we haven’t removed the moral dimension. We’ve just hidden it inside an optimisation function that nobody is examining. 6
The connected vehicle surveillance apparatus — where telematics data flows from your car to data brokers and then to insurers without meaningful consent — is a concrete example of what Magnifica Humanitas identifies as the “new monopolies of AI.” The encyclical expresses explicit concern about these new monopolies, warning that concentration of data and algorithmic power in the hands of a few entities constitutes a structural threat to the common good — regardless of whether the entities involved have malicious intentions. The GM OnStar scandal, in which driving data from millions of vehicles was sold to LexisNexis and then to insurers who used it to raise premiums without customer knowledge, is not a story about bad actors. It’s a story about a system operating exactly as its incentives dictated — and producing outcomes that no reasonable moral framework could endorse. The full scope of that scandal and what it means for every connected vehicle owner is detailed in our piece on telematics insurance privacy risks and what your car is really tracking.7
The Technocratic Paradigm and the Insurance Economy
Magnifica Humanitas uses the phrase “technocratic paradigm” to describe a way of thinking — and ultimately a way of organising society — in which technical efficiency becomes the primary measure of value. In this paradigm, the question is not “is this right?” but “does this optimise the metric?” The insurance industry has been moving steadily toward this paradigm for the better part of a decade, and the shift has produced genuine efficiency gains alongside genuine harms that the efficiency framing makes structurally difficult to see.
An October 2025 review and meta-analysis in the British Medical Bulletin found that AI chatbots were rated as showing more empathy than human healthcare professionals in 13 of 15 studies that compared them. That finding is genuinely interesting and genuinely complicated. An AI that produces empathic-sounding responses may be more consistently responsive than an overworked human professional — and yet something is categorically different about an entity that simulates empathy through pattern matching and one that feels it through shared vulnerability. Magnifica Humanitas addresses this directly in its concern about human relationships being replaced rather than enhanced — not as sentimentality, but as a substantive claim about what makes care meaningful. 8
In the insurance context, the “care” dimension is often invisible because the industry rarely presents itself in those terms. But insurance, at its social foundation, is a solidarity mechanism: people pool resources against shared risk, and the system works because of mutual obligation. Human-centred AI research distinguishes between augmentation — using AI to enhance human capability while preserving human agency — and automation, which replaces human judgment entirely. A major aim of human-centred AI is toward automating rote tasks that hinder human productivity, freeing people to direct energy toward higher-level tasks, thus achieving augmentation through automation rather than replacement. The insurance industry has, in many deployments, skipped the distinction entirely — replacing underwriters with models, claims adjusters with algorithms, and customer service representatives with chatbots, without asking whether the human judgment being eliminated carried value that wasn’t captured in the efficiency metrics. 9
The embedded insurance model — where coverage is sold through digital checkout flows, with AI pricing in real time from behavioural and transactional data — is perhaps the purest expression of the technocratic paradigm applied to financial protection. It is frictionless, personalised, scalable, and almost entirely opaque to the person being served. Our analysis of embedded insurance and the hidden coverage inside apps and checkout pages maps the data infrastructure underneath that frictionlessness — and the question Magnifica Humanitas forces is whether frictionless protection is the same as genuine protection, or whether something essential is lost when a moral act of mutual solidarity is reduced to a checkout toggle.
Social Justice, AI, and the Right to Be Treated as a Person
The section of Magnifica Humanitas that is most directly applicable to the AI insurance economy is its articulation of social justice principles as a framework for evaluating algorithmic systems. The encyclical’s 245 paragraphs include a dedicated treatment of subsidiarity, solidarity, and the universal destination of goods — classical social doctrine principles — applied to AI development and governance. The Pope argues that AI governance cannot be left to market forces or even to individual national regulatory frameworks, but requires “shared standards of social justice” that cross jurisdictions and reflect the common good rather than the interests of the data-richest actors. 10
This has a direct implication for how we think about the fragmented regulatory patchwork currently governing AI in insurance — where Colorado has one standard, New York another, the EU a third, and much of the world none at all. Antiqua et Nova concluded that artificial intelligence must only be used to complement human intelligence rather than replacing it, as replacement would enslave humanity and serve as a substitute for genuine human relations. The subsidiarity principle — that decisions should be made at the most local level capable of making them well — does not mean that insurance regulators in Wyoming can adequately govern the AI models deployed by multinational insurers. It means that governance must be designed at a scale appropriate to the power of the system being governed. 11
This has a direct implication for how we think about the fragmented regulatory patchwork currently governing AI in insurance — where Colorado has one standard, New York another, the EU a third, and much of the world none at all. Antiqua et Nova concluded that artificial intelligence must only be used to complement human intelligence rather than replacing it, as replacement would enslave humanity and serve as a substitute for genuine human relations. The subsidiarity principle — that decisions should be made at the most local level capable of making them well — does not mean that insurance regulators in Wyoming can adequately govern the AI models deployed by multinational insurers. It means that governance must be designed at a scale appropriate to the power of the system being governed. 12
The liability question that sits at the heart of our AI liability insurance analysis — who pays when algorithms make expensive mistakes — is, in the framework of Magnifica Humanitas, ultimately a question about accountability before persons, not just accountability before law. When a machine learning model denies a claim that a human adjuster would have approved, the question isn’t only “which entity owes damages?” It’s “who was responsible for designing a system that treated this person as a data point rather than a human being with a legitimate claim?” That’s a moral question that the legal frameworks are only just beginning to approximate.
For workers and contractors navigating AI-augmented work environments — where their cyber risk profile is assessed by models they cannot inspect, their coverage may exclude AI-generated harms, and their insurer’s own claims systems may be operated by the same category of algorithm that prices them — the cyber insurance landscape for remote workers in 2026 sits inside the same broader question Magnifica Humanitas is asking. Are these systems serving the people they ostensibly protect, or are they optimising a different objective entirely?
Frequently Asked Questions
Magnifica Humanitas — Latin for “Magnificent Humanity” — is Pope Leo XIV’s first encyclical, signed on May 15, 2026 and released May 25, 2026. Titled “On Safeguarding the Human Person in the Time of Artificial Intelligence,” it is a 245-paragraph formal teaching document positioning AI as the defining social challenge of the present era, equivalent to industrialisation in 1891. The document calls for shared standards of social justice in AI development, warns against the concentration of AI power in the hands of a few, and argues that technology is never morally neutral — it takes on the characteristics of those who design, fund, and regulate it. It matters because it represents the most comprehensive moral framework yet offered by a major global institution for evaluating AI governance, going beyond regulatory compliance to ask fundamental questions about human dignity and the common good.
While the encyclical does not address insurance specifically, its principles apply directly. It condemns “new monopolies of AI” — the concentration of data and algorithmic power in corporate hands — and calls for AI to complement rather than replace human judgment, especially in fields where “the fragility of the human condition” is most evident. Healthcare and insurance are precisely those fields. The document insists that those responsible for deploying AI in consequential domains carry the same moral obligation as those who traditionally provided care — and that efficiency optimisation does not override the duty to treat each person as an individual of irreducible dignity.
The encyclical’s warning that “AI tends to amplify the power of those who already possess economic resources, expertise and access to data” is a precise description of algorithmic bias in insurance. Models trained on historically biased data — where communities of colour were denied coverage or overcharged through redlining and discriminatory underwriting — reproduce those patterns at computational scale through proxy variables like ZIP codes and credit scores. Magnifica Humanitas frames this not merely as a legal or regulatory problem but as a structural injustice requiring shared standards of social justice, not just individual market corrections.
Antiqua et Nova (“Ancient and New”) is a 2025 doctrinal note issued by the Vatican’s Dicastery for the Doctrine of the Faith under Pope Francis. Its approximately 30 pages contrast humanity’s relational, truth-seeking nature with AI systems that operate through pattern recognition and lack the creative, spiritual, and moral dimensions of human thought. It concluded that AI should only complement, not replace, human intelligence. Magnifica Humanitas builds directly on Antiqua et Nova, citing it throughout as groundwork for its broader social doctrine application — treating the earlier document as the theological foundation for the practical governance principles Leo XIV now develops.
Yes, significantly. The encyclical calls for governance structures that cross national jurisdictions and reflect the common good rather than the interests of the most data-rich actors — a direct challenge to the fragmented state-by-state regulatory patchwork currently governing AI in insurance. It applies the classical social doctrine principle of subsidiarity, arguing that governance must operate at a scale appropriate to the power of the system being governed. For AI models deployed by multinational insurers, that means international coordination, not just domestic compliance — a standard that current regulatory frameworks in most jurisdictions fall well short of.
The Bottom Line: Magnificent Humanity in the Age of the Black Box
Magnifica Humanitas arrived on May 25, 2026, the same week that the insurance industry’s AI adoption continues its 25% annual spending growth trajectory, that class action suits over algorithmic claim denial are moving through discovery, and that regulators in Colorado, New York, and Brussels are building governance frameworks against a clock set by the pace of deployment rather than the pace of deliberation. The Pope’s timing was not accidental.
The encyclical identifies a deeper danger than the familiar concerns about job displacement and privacy violations: that human beings may begin to see themselves and others as purely processable data. In the insurance context, that danger is not abstract. The person whose health claim was denied by a model that reviewed 300,000 applications in two months is already being processed rather than assessed. The driver whose premium rose 21% because their connected car sold their data to a broker without their knowledge is already being valued as a data asset rather than respected as a person. The community overcharged for home insurance because its ZIP code encodes a history of segregation is already being subjected to mathematics that masquerades as objectivity. 13
The encyclical’s foundational premise — that technology is never neutral because it takes on the characteristics of those who devise, finance, regulate, and use it — is the moral lens through which every deployment of AI in consequential domains should be evaluated. Not just: does this model perform accurately? But: whose interests does this system serve? Who bears the cost of its errors? What happens to the person at the end of its output? And is there a human being who can be held accountable for that outcome? 14
External Links
All the best
Enjoy every single day