Magnifica Humanitas and MANAV Share Seven Principles for AI

In early 2026, two governing frameworks for artificial intelligence arrived from opposite ends of the world’s civilizational map. Pope Leo XIV’s Magnifica Humanitas (a 42,300-word encyclical signed May 15 and published May 25 by the Holy See) and Prime Minister Narendra Modi’s MANAV framework (unveiled at the India AI Impact Summit in New Delhi in February) share seven structural principles on dignity, inequality, data, algorithmic bias, labor, and global governance. Neither document cites the other.

That convergence has a precise historical parallel: a 1947 moment when a UNESCO committee on the philosophical foundations of human rights found that radically different traditions agreed on the same core commitments about the human person. What happened next, and how long it took to produce anything enforceable, is the harder part of the story.

The 1948 Precedent

Before the United Nations drafted its declaration on human rights, UNESCO convened a committee of philosophers and political scientists with one specific task: find out whether Confucian, Islamic, Hindu, Catholic, and secular liberal traditions actually agreed on anything about the human person. The committee surveyed thinkers from across those traditions, including Mohandas Gandhi and Aldous Huxley. The main conclusion: despite profound doctrinal differences, the traditions shared a core commitment to a life free from systematic degradation and insecurity.

That philosophical canvassing directly informed the drafting of the Universal Declaration of Human Rights, adopted by the UN General Assembly on December 10, 1948. Forty-eight member states voted in favor, none against. The drafting committee had been assembled from representatives with vastly different political, cultural, and religious backgrounds; their debates over the meaning of human dignity produced 30 articles they could all ratify. The UDHR has since inspired more than 70 binding human rights treaties and been translated into over 500 languages, making it the most translated document in UN history.

None of that happened quickly. The UDHR itself carries no enforcement mechanism; its power rests on voluntary state alignment with its principles and on the slow work of treaty negotiation that followed over decades.

Neither framework was coordinated with the other. Together they have produced something close to what the 1947 UNESCO survey found: shared conclusions from incompatible starting points about a specific challenge to human welfare. The question is whether that convergence travels the same route, from moral consensus to shared vocabulary to something enforceable, and at the pace that artificial intelligence demands.

Seven Places They Agree

The seven convergences span the most contested questions in AI governance: who the technology serves, who it harms, and who controls it. The table below draws from the full encyclical text on the Vatican’s website and from the Prime Minister’s post-summit article on the MANAV vision.

Among the seven convergences, the one on data and personhood has the most direct language in both documents. The encyclical argues against “the pretence that a single language, even a digital one, can translate everything, including the mystery of the person, into data and performance.” In his post-summit article, the Prime Minister wrote that “humans must never become mere data points or raw material for machines.” The two sentences, drafted without coordination, reach the same position: that an algorithm’s data cannot define a human being.

A Framework Grounded Differently

The seven convergences coexist with real structural differences. Their foundations and emphases diverge in ways that matter for how each translates into practice.

The encyclical is theological at its root. It situates AI within a tradition of Catholic social teaching running from Rerum Novarum, Pope Leo XIII’s 1891 document on industrial labor, through Laudato Si’, Pope Francis’s 2015 environmental encyclical. Its test for AI is grounded in what it calls human dignity as a gift of creation. The document also asks the Church to conduct an “examination of conscience” about its own institutional power abuses, including “spiritual, economic, institutional, sexual and power-based abuse.” That self-critical move has no equivalent in the framework the Prime Minister presented.

The framework is a policy acronym. Its five pillars are M (Moral and Ethical Systems), A (Accountable Governance), N (National Sovereignty), A (Accessible and Inclusive technology), and V (Valid and Legitimate systems). These pillars are built for operationalization in regulation and institutional practice. The “N” pillar asserts that data generated by citizens remains under national jurisdiction. “Whose data, his right,” the Prime Minister said at the February summit. That framing has no analog in the encyclical, which argues from universal human dignity rather than state sovereignty.

This gap isn’t merely philosophical. A governance framework built on national sovereignty and one built on universal personhood will reach different answers to the same concrete question: whose rules apply when an AI system trained in one country operates in another?

Global South Stakes

The India AI Impact Summit in New Delhi was the first in the global AI summit series hosted by a nation outside the Western axis. Both frameworks treat this geographic fact as load-bearing.

The Vatican speaks to roughly 1.4 billion Catholics worldwide, a large majority of them in Africa, Latin America, and Asia. When the encyclical warns about AI decisions concentrated in too few hands, that warning lands differently in Lagos or Manila than in San Francisco or Brussels. The document was published simultaneously in eight languages, with no Latin version required, the first encyclical released that way.

• 100+ countries represented at the New Delhi summit
• 20+ heads of state attended
• 250,000 citizens pledged ethical AI use at the summit, a Guinness World Record
• 8 languages in which the encyclical was simultaneously published

Both frameworks explicitly target the same structural failure: AI systems trained predominantly on Western data, presented as universal tools. The encyclical describes the mechanism as bias “cloaked in a veneer of neutrality and objectivity.” The Prime Minister’s formulation is more direct: AI trained in a “Western context may not do justice to our diversity, linguistic, cultural and regional.” The encyclical’s claim is theological, that human personhood exceeds any data representation. From a more geopolitical vantage, the Prime Minister cites training data skewed toward Western languages that systematically disadvantages populations outside that context.

The summit’s scale reflected the appetite for a broader conversation. UN Secretary-General António Guterres, speaking in New Delhi, said the future of AI “cannot be decided by a handful of countries.” India also used the summit to notify the IT Amendment Rules 2026, formally defining “Synthetically Generated Information” and mandating a three-hour takedown window for illegal AI-generated content, a domestic regulation that took legal effect without waiting for any global framework.

From Consensus to Enforceable Rules

The New Delhi Declaration and the Frontier AI Commitments signed by tech companies at the summit are explicitly voluntary. The encyclical carries no legal authority over any AI developer or government. Major tech companies pledged to publish anonymized usage data and evaluate systems for global deployment contexts; as of June 2026, that pledge is the only documented mechanism connecting either framework to actual AI development practice.

At the Vatican presentation of the encyclical on May 25, Anthropic co-founder Chris Olah spoke alongside Professor Anna Rowlands and Cardinals Víctor Manuel Fernández and Michael Czerny. Olah framed the relationship between AI builders and moral institutions plainly:

A long collaboration between those of us who are building this and those who can see what we, from inside, cannot.

Olah, co-founder of Anthropic, at the Vatican, May 25, 2026.

The 1947 UNESCO survey generated a shared vocabulary that made the UDHR politically possible. The UDHR informed decades of treaty negotiation. The International Covenant on Civil and Political Rights came in 1966, eighteen years after the declaration. AI systems are now shipping major capability updates faster than an eighteen-month legislative cycle.

As of June 2026, there is no international AI treaty. The two frameworks that agree most on what AI governance should require are both, by design, non-binding.

Frequently Asked Questions

What does MANAV stand for in Modi’s AI framework?

MANAV is an acronym for five governance pillars unveiled by the Prime Minister at the New Delhi summit in February 2026: Moral and Ethical Systems, Accountable Governance, National Sovereignty, Accessible and Inclusive technology, and Valid and Legitimate systems. The word “MANAV” itself means “human” in Hindi and Sanskrit, a point the Prime Minister cited explicitly: “MANAV means human.”

What is Magnifica Humanitas about?

The document, whose title translates as “Magnificent Humanity,” is the current pontiff’s first encyclical, signed May 15, 2026 and published May 25, 2026. The 42,300-word text, divided into five chapters, addresses how society must safeguard human dignity in the age of artificial intelligence, covering labor rights, truth, freedom, social justice, algorithmic bias, AI in warfare, and the risks of reducing persons to data. The Vatican published it in eight languages simultaneously; it was the first encyclical released without an official Latin version.

How does Magnifica Humanitas relate to Rerum Novarum?

Leo XIV signed the encyclical on May 15, 2026, the 135th anniversary of Pope Leo XIII’s Rerum Novarum (1891), the foundational social encyclical addressing labor rights during the Industrial Revolution. Leo XIV chose his papal name in direct reference to his predecessor and described AI as a “new industrial revolution” requiring a comparable moral response. The document builds on a chain of social teaching including Laudato Si’ (Pope Francis, 2015) and Caritas in Veritate (Pope Benedict XVI, 2009).

Is the MANAV framework legally binding?

No. The MANAV vision and the New Delhi Declaration signed at the India AI Impact Summit 2026 are voluntary commitments. Tech companies attending pledged to publish anonymized usage data to help developing nations track AI’s economic impact, but no treaty mechanism enforces compliance. India separately enacted the IT Amendment Rules 2026, which carry domestic legal force: they formally define “Synthetically Generated Information” and require takedown of illegal AI-generated content within three hours.

Who attended the Vatican presentation of Magnifica Humanitas?

The Pope presented the encyclical personally at the Vatican on May 25, 2026, alongside Anthropic co-founder Chris Olah, Professor Anna Rowlands, Cardinal Víctor Manuel Fernández, Cardinal Michael Czerny, and Professor Léocadie Lushombo. A reigning pope presenting an encyclical personally was unusual; most delegate the task to cardinals. Olah praised the Vatican’s role as “informed critics” of AI development and described the event as the start of a “long collaboration.”

What was the India AI Impact Summit 2026?

The summit was held in New Delhi in February 2026, making it the first in the global AI summit series hosted by a Global South nation. Representatives from more than 100 countries attended, including over 20 heads of state. The Prime Minister unveiled the framework at the summit. Major tech companies signed the Frontier AI Commitments, a voluntary pledge to evaluate AI systems for global deployment contexts. The summit set a Guinness World Record: over 250,000 citizens pledged to use AI ethically within a 24-hour window.

What is the Global South’s stake in AI governance?

Both frameworks explicitly flag Western dominance of AI development as a structural problem. The encyclical warns of bias “cloaked in a veneer of neutrality,” while the framework’s Moral Systems pillar targets algorithmic guardrails for socio-cultural bias. India is the world’s most populous country and one of the fastest-growing digital markets; the Vatican reaches roughly 1.4 billion Catholics, the majority of whom live outside Western Europe and North America. Both institutions argue that AI governance shaped primarily by the United States, China, and the European Union will systematically disadvantage populations those frameworks weren’t designed to serve.