The Pope vs. the Algorithm: Inside the Vatican's Bid to Make AI Ethics a Matter of Faith

On May 25, 2026, Pope Leo XIV became the first pope in history to personally present an encyclical to the world. In the Vatican's Synod Hall, seated among a row of cardinals and theologians, was an unusual guest: Chris Olah, co-founder of the AI company Anthropic [1]. The document they were there to unveil — Magnifica Humanitas ("Magnificent Humanity") — runs 43,000 words and represents the most ambitious attempt by any religious institution to claim moral authority over the development of artificial intelligence [2].

Leo XIV signed the encyclical on May 15, 2026, exactly 135 years to the day after his namesake, Pope Leo XIII, signed Rerum Novarum, the 1891 encyclical that reshaped Catholic social teaching on labor rights during the Industrial Revolution [3]. The parallel is deliberate. Where Leo XIII confronted the exploitation of workers by industrial capitalism, Leo XIV frames AI as posing "even greater consequences" than the upheaval of the 19th century [4].

What the Encyclical Actually Says

Magnifica Humanitas is addressed "to all the Catholic faithful, to all Christians and to men and women of goodwill" [5]. Its core argument: artificial intelligence "cannot be treated as morally neutral," and decisions about its development and deployment are matters of faith, not merely policy [6].

The encyclical makes several specific claims and demands:

On autonomous weapons, the pope states that some weapons systems have advanced "practically beyond any human reach to govern them" and calls for their "disarmament" [7]. The document cites the wars in Ukraine, Gaza, Lebanon, and Iran as evidence of "the inhuman evolution of the relationship between war and new technologies in a spiral of annihilation" [8]. Decisions that end human life, it argues, "cannot be delegated to systems incapable of understanding moral weight" [8].

On algorithmic decision-making, the pope warned of "very troubling" systems and algorithms "capable of denying access to healthcare, employment or security based on unjust and prejudiced data" [7]. The encyclical frames AI as already "deciding who gets hired and who gets fired, whether you receive credit or are denied a loan, what your face means to a surveillance camera" [7].

On generative AI, the document warns that when machines simulate "human voices and faces, wisdom and knowledge, consciousness and responsibility, empathy and friendship," they encroach upon human relationship and communication itself [8].

On labor, Leo XIV decried "new forms of slavery" endured by workers who tend AI systems and factory workers producing the devices on which AI runs [7]. He warned that workers "often must adapt to machine speeds rather than machines supporting human workers" [9].

On transhumanism, the pope directly critiques philosophies that interpret "progress as surpassing the human condition," positioning them as incompatible with Catholic teaching on human dignity [9].

Doctrinal Weight: Where Does This Fit?

An encyclical is among the highest forms of papal teaching, though it is not an ex cathedra pronouncement (an infallible declaration on faith or morals). Under Catholic doctrine, the faithful are called to receive the pope's ordinary teaching with "religious assent" — a standard that falls short of absolute obedience but carries significant moral weight [5].

Kathleen Sprows Cummings, a historian at the University of Notre Dame, described Magnifica Humanitas as "more humble in tone than Rerum Novarum" yet "far more capacious" in scope [6]. The comparison to Rerum Novarum is instructive: that 1891 document never became law anywhere, yet it shaped a century of labor policy, supported the right of workers to form trade unions, rejected both state socialism and laissez-faire capitalism, and influenced political movements across Europe and Latin America [10].

Laudato Si', Pope Francis's 2015 encyclical on the environment, offers a more recent precedent for measuring impact. Published months before the Paris climate conference, it helped mobilize 900,000 Catholics to sign a climate petition calling for a 1.5°C warming target and 40,000 Catholics to join Global Climate March rallies [11]. The Vatican received credit — alongside island nations and other faith groups — for the Paris Agreement's inclusion of the 1.5°C aspirational target [11]. An editorial in Nature praised the encyclical for its statements on sustainability and global poverty [12].

Whether Magnifica Humanitas will produce comparable policy momentum remains an open question. The encyclical calls for stricter state and international regulations on AI companies, transparent benefit-sharing with communities, tax systems placing greater burden on resource-rich entities, international laws addressing automated weapons, and individual control over personal data [9].

The Anthropic Question

The most scrutinized aspect of the encyclical's release was the presence of Chris Olah at the Vatican presentation. A Vatican source told reporters that Anthropic's inclusion "is not an endorsement, prize, reward or canonization" [3]. But the optics of an AI company co-founder seated among cardinals at the launch of a document calling for AI regulation raised immediate questions.

The relationship has context. Anthropic came into conflict with the Trump administration after refusing to loosen safeguards preventing its AI models from being used for lethal autonomous warfare without human oversight or for mass surveillance [3]. The Pentagon labeled Anthropic a "supply chain risk" — the first time such a designation was applied to a U.S. company [3]. At the Vatican event, Olah acknowledged that AI development "operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing" and called for broader engagement from religious communities [9].

This is not the Vatican's first entanglement with the tech industry. In February 2020, the Pontifical Academy for Life launched the Rome Call for AI Ethics, a document signed by Microsoft, IBM, the FAO, and the Italian government [13]. The Call established six principles — transparency, inclusion, responsibility, impartiality, reliability, and security — and has since expanded from five original signatories to include Jewish and Muslim religious leaders (2023), government representatives (2024), and a broader range of corporate participants [14].

Source: Rome Call for AI Ethics

Data as of May 1, 2026CSV

The tension is real: the same companies named as potential risks in papal documents are also partners in Vatican AI initiatives. Whether this represents effective engagement or compromised moral authority depends on whom you ask.

The Theological Debate Within

Catholic teaching distinguishes between binding moral principles and prudential judgments. As Catholic commentator Jeff Mirus has written, "The moral principles at work in Catholic social teaching are binding on all. The prudential judgments are binding on nobody" [15]. This distinction matters because much of Magnifica Humanitas operates in the space between universal moral claims (human dignity is inviolable) and specific policy recommendations (how to regulate algorithmic hiring).

Cardinal Víctor Manuel Fernández, prefect of the Vatican's Dicastery for the Doctrine of the Faith, defended the encyclical at its presentation by arguing that Catholic teaching holds "every human being has infinite dignity" — a principle he positioned as directly relevant to AI governance [16]. Meghan Sullivan, a philosopher at Notre Dame, praised it as "one of the most compelling treatments of AI ethics" she had encountered, emphasizing its practical guidance for corporate leaders, policymakers, and educators [6].

But the question of whether AI ethics falls within the Church's doctrinal competence — or represents prudential judgment that individual Catholics may legitimately dispute — has not been fully resolved. Some theologians warn that "if AI replaces prudential judgment, prudence weakens in the human person," while others argue that specific technology policy recommendations lie primarily within the laity's province [17]. The encyclical itself appears aware of this tension, framing its arguments through the lens of human dignity and imago Dei (the belief that humans are made in God's image) rather than issuing canonical penalties for non-compliance [6].

No cardinals or bishops have publicly broken with the pope over the document in its first days. But the Catholic commentariat is already parsing which of its claims carry the weight of moral teaching and which are policy suggestions open to disagreement.

The Institutional Ripple Effect

The Catholic Church operates one of the largest institutional networks on earth. Its hospitals, universities, schools, and social service centers span every continent.

Source: Center for Applied Research in the Apostolate (CARA)

Data as of Jan 1, 2025CSV

These institutions buy software, set procurement standards, educate students, and shape public expectations [18]. An encyclical touching AI governance creates practical questions at every level: Can a Catholic hospital use algorithmic triage systems? Must a Catholic university restrict its use of generative AI tools? How should Catholic schools approach AI in education?

The encyclical does not issue specific compliance mandates. But as the Axios headline put it, Pope Leo has set "Catholics on collision course with AI" [18]. Encyclicals are read carefully by Catholic-affiliated institutions, and bishops, university presidents, and hospital administrators in Catholic-majority regions — Italy, Spain, Poland, Ireland, the Philippines, much of Latin America — will face pressure to align purchasing and governance decisions with the document's principles [18].

The Catholic Health Association of the United States has already been examining AI's ethical implications in healthcare delivery [18]. For institutions caught between the encyclical's moral framework and national regulations or employer mandates that require AI adoption, the path forward is unclear.

How Other Faiths Approach AI

The Catholic Church is not the only religious tradition grappling with AI ethics, but Magnifica Humanitas represents the most institutionally authoritative statement any faith has produced.

In January 2023, Jewish and Muslim leaders formally joined the Rome Call for AI Ethics at the Vatican [14]. Chief Rabbi Eliezer Simha Weisz of the Council of the Chief Rabbinate of Israel and Sheikh Al Mahfoudh Bin Bayyah, Secretary General of the Abu Dhabi Forum for Peace, signed on behalf of their communities [14]. In July 2024, the Rome Call expanded further in Hiroshima, with additional world religious leaders committing to its principles [19].

Islamic ethics emphasizes khilafah (stewardship) over the earth and frames responsible technology use as a religious obligation [20]. Judaism's concept of tikkun olam (repairing the world) encourages innovation when aligned with justice and fairness, while scholars caution against AI replacing human connection or diminishing the sanctity of life [20]. Protestant and evangelical traditions emphasize responsible stewardship over creation, though they lack a centralized authority equivalent to the papacy to issue binding ethical pronouncements [20].

The Catholic framing of human dignity through imago Dei — the idea that every human bears the image of God and therefore possesses irreducible worth that no algorithm can assess — is broadly compatible with these other traditions, even if the specific theological vocabulary differs. The practical question is whether a unified interfaith voice on AI can produce policy results that individual traditions cannot.

The Academic Explosion

The encyclical arrives amid an extraordinary surge in academic attention to AI ethics. Research publications on "artificial intelligence ethics" have grown from roughly 1,200 papers in 2011 to nearly 117,000 in 2025, according to OpenAlex data [21].

Source: OpenAlex

Data as of Jan 1, 2026CSV

This volume of scholarship suggests that the Vatican is entering a conversation already well underway in secular institutions. Whether a papal encyclical can redirect or reshape that discourse — or whether it will be absorbed as one voice among many — depends partly on how seriously Catholic universities and research institutions translate its principles into concrete governance frameworks.

Nitesh Chawla, a computer scientist at Notre Dame, highlighted the encyclical's insistence that moral and technical questions "cannot be separated" [6]. Paolo Carozza, of Notre Dame's law school, called it "a defining document for our era" but emphasized that its core question is practical: does AI "serve human development or enable exploitation?" [6].

What Comes Next

Pope Leo XIV has already created an interdicasterial commission on AI — a cross-departmental Vatican body tasked with ongoing study and guidance [22]. The commission signals that Magnifica Humanitas is intended as a beginning, not a final word.

The document's real test will come in the months and years ahead, measured by whether Catholic institutions change their technology procurement, whether Catholic-majority nations incorporate its principles into legislation, and whether the Vatican's partnership model with companies like Anthropic produces meaningful guardrails or merely photo opportunities.

The historical record is mixed. Rerum Novarum shaped labor movements for a century. Laudato Si' contributed to the political pressure behind the Paris Agreement but could not prevent nations from falling short of their commitments [11]. Magnifica Humanitas arrives at a moment when AI development is accelerating faster than any regulatory framework — secular or religious — can keep pace.

As Chris Olah said at the Vatican presentation, AI development operates inside incentives that "can sometimes conflict with doing the right thing" [9]. The pope's answer is that doing the right thing is not optional — it is, he argues, a religious obligation. Whether 1.4 billion Catholics, and the institutions that serve them, will treat it that way is the question the encyclical cannot answer on its own.