Vatican Certifies Pope Leo XIV's AI Encyclical as Human-Written

Pope Leo XIV wrote an extensive document warning against the uncontrolled development of artificial intelligence. A few weeks later, he had to explain himself over whether he had used the very tools he warns about while writing it. The Vatican commissioned an outside firm to formally certify the encyclical's authorship, the first such case in the Church's history.

Accusations Online

"Magnifica Humanitas" is the first document of this rank in the Church's history devoted entirely to artificial intelligence. In it, Leo XIV writes about the need to protect human dignity in the era of advanced algorithms, calls for regulation, and warns against treating AI as a tool for advancing the particular interests of a narrow group of companies.

Soon after publication, voices online suggested the document itself may have been produced with the help of generative artificial intelligence. The discussion was sparked by researcher Lynch Zhang, writing for The Verge, who pointed to passages he said carried statistical hallmarks characteristic of text generated by language models. Zhang, however, explicitly stated that his observation was not proof, only a hypothesis worth checking.

How the Vatican Checked

Rather than dismiss the accusations with silence, the Vatican took an unprecedented step: it commissioned an outside firm to formally certify the document's authorship. The choice fell on the Australian company Proudly Human, led by Dr. Alan Finkel, Australia's former Chief Scientist. The procedure included comparing the encyclical's style with the pope's earlier speeches and writings, and required signing a formal declaration that artificial intelligence had not been used in the writing process.

The verification method itself relies on a set of several tools analyzing different features of writing style and grammar, whose results the company combines in its own scoring algorithm. Among the detectors used is Pangram, one of the more popular tools for detecting text generated by language models. Commentators have pointed out the paradox of the situation: the very analysis meant to prove that a document about the dangers of AI was not produced with AI relies on artificial intelligence algorithms.

The encyclical is 94 percent human-written, 4 percent AI-generated, and another 2 percent classified as AI-assisted content.

What the Pope Wrote

The document itself, dated to mid-May, refers directly to Leo XIII's social encyclical from more than a hundred years ago and touches on several themes at once. Leo XIV calls for neither demonizing digital tools nor deifying them, but consciously guiding them. The document contains a passage that has become one of the most widely quoted: the pope writes that technology in itself is neither good nor bad, and it is humans who give it direction.

The first task before us is not to demonize the tools, nor to deify them, but to guide them, starting from one fundamental principle: truth is a common good. - Leo XIV, encyclical Magnifica Humanitas

The encyclical places particular emphasis on protecting children from violence and disinformation generated by algorithms, as well as on the growing problem of loneliness, which the Vatican says is deepened by relationships with chatbots replacing human contact. The pope also calls for regulating the market and breaking the monopolies of the few companies controlling the development of the most advanced models.

Why It Matters

The case shows how deeply suspicion of AI use has already penetrated the assessment of the credibility of public texts, even those of the highest institutional rank. If AI-text detectors are a reliable enough argument that the Vatican itself turns to them to defend the authenticity of a papal document, similar tools are likely to be used increasingly often in disputes over the authorship of academic papers, news articles, or official documents.

On the other hand, the method itself remains controversial. AI-text detectors, including Pangram, used in this case, have repeatedly been criticized for errors, especially false positives against text written by non-native speakers of a given language or by people writing in the formalized, repetitive style typical of Church documents.

For the Vatican, which uses the encyclical to call for caution toward artificial intelligence, the certification result carries reputational weight. An institution criticizing the uncontrolled development of technology could not afford accusations that it secretly used that very technology on the pontificate's most important doctrinal document.