(LifeSiteNews) — An independent researcher has alleged that parts of Pope Leo XIV’s recent encyclical addressing artificial intelligence (AI) have, in keeping with the theme of the document, been written by AI.
Linch Zhang, in a Substack piece published on Tuesday, said that the AI detector Pangram flagged portions of Pope Leo XIV’s encyclical Magnifica Humanitas, “On Safeguarding the Human Person in the Time of Artificial Intelligence,” as written by AI.
“Pangram is by far the best commercially available AI detector,” claimed Zhang. “It is much better than other AI detectors, so much so that other ones are almost useless in comparison. In particular, Pangram optimizes very hard for getting a false positive rate of nearly zero, while being more okay with false negatives,” he explained.
According to Zhang, Pangram found that “some paragraphs” of Magnifica Humanitas are between 40 percent and 100 percent AI-generated, while most paragraphs are written by humans.
For example, paragraphs seven and eight were flagged by Pangram as “fully AI-generated,” while paragraphs 122 and 123 registered as 60 percent fully AI-generated.
“This indicates to me that some senior Vatican officials who contributed to the encyclical used AI assistance heavily and most (probably including Pope Leo himself) did not,” commented Zhang.
In fact, he has hypothesized that Pope Leo does not endorse the use of AI in encyclicals “and plausibly was not even aware of significant AI usage in his own encyclical!”
For comparison, Zhang used Pangram to assess the last four papal encyclicals, all written by Pope Francis. He found that “the first 20 paragraphs on all of them register as 100% human, all with high confidence.” Encyclicals issued by Pope Benedict XVI and Pope John Paul II likewise registered as fully composed by humans.
In addition, he ran Pope Leo’s recent speech announcing the AI encyclical through Pangram and found that it was assessed to be “100%” written by humans. “This is evidence that Pope Leo himself and/or his primary speechwriter does not use AI to draft his speeches,” said Zhang.
Zhang said that he reads AI-generated text as part of his job regularly and that his familiarity with AI writing, particularly with the Claude AI model, has sensitized him to its signs. He noted these include the frequency of em-dashes, the generous use of the word “genuinely,” often used by the AI model Claude, as well as the dense use of “tricolons,” which are “a series of three parallel words, phrases, or clauses used for rhetorical effect.
The researcher pointed out that these indicators can be part of writing-style quirks, stressing the importance of holistic analysis, in particular, that provided by a quality AI-detector such as Pangram.
Zhang believes he has found the “voice” of the AI model Claude in the recent encyclical about “safeguarding the human person in the age of artificial intelligence,” a feature he found to be “ironic.”
Zhang is the founder of Open Asteroid Impact, a satirical, fictional startup “dedicated to intentionally lobbing asteroids at Earth for mineral rights.” Intended to parody Open AI, the “startup’s” imaginary mission is to hurl asteroids to Earth “for mineral rights” in the name of safety, “arguing that if they (we) don’t accelerate as quickly as possible, other, more dangerous, competitors would do so before us.”
Many figures have made a version of this argument while arguing for U.S. AI development, such as Sam Altman, who has said the U.S. “absolutely has to win” the AI race. Vice President J.D. Vance said at the Paris AI Summit last year, “I’m not here to talk about AI safety … I’m here to talk about AI opportunity.” He said he believes that when it comes to AI, “our tendency is to be too risk averse.”
“We believe that excessive regulation in the AI sector could kill a transformative industry just as it’s taking off,” said Vance at the time.
Now that Americans are increasingly concerned about the dangers AI poses, such as its threat to jobs, intellectual development, and mental health, the Trump administration is reportedly considering creating an AI oversight group.