Did Pope use AI to write his anti-AI encyclical? The internet is debating

Pope Leo XIV has barely released his first major encyclical, and the internet is already asking an awkwardly modern question: did artificial intelligence help write the Vatican’s warning about the dangers of artificial intelligence?

The lengthy document, titled Magnifica Humanitas, focuses on the social, ethical and political risks posed by rapidly advancing AI systems. In it, Pope Leo urges governments and technology companies to slow down the race towards increasingly powerful AI tools, warning that unchecked development could fuel misinformation, deepen conflict and threaten workers’ rights.

But while the document’s message was about caution, online discussion quickly shifted towards the possibility that the encyclical itself may have been drafted with AI assistance.

Did Pope use AI to write dangers of AI?

The debate began after researcher Linch Zhang posted an analysis on the online forum LessWrong, claiming parts of the encyclical showed signs commonly associated with AI-generated writing.

More from Tech

Using the AI detection tool Pangram, Zhang found that certain passages of the document were estimated to be anywhere between 40 per cent and fully AI-generated. Another user reportedly analysed the text chapter by chapter and found that nearly two-thirds of the first chapter was flagged as AI-written.

Even independent tests produced mixed results. When portions of the document were run through Pangram, some sections appeared strongly human-written, while others showed noticeable AI markers.

One of the clues pointed out by Zhang was the repeated use of the word “genuinely”, which researchers have previously linked to writing patterns produced by Anthropic’s Claude models.

Still, none of this proves that AI actually wrote the encyclical.

AI detection tools remain controversial and are far from perfect. Different systems can produce wildly different results depending on the text, formatting and writing style involved. Even companies behind these tools acknowledge limitations.

Pangram, however, has earned some credibility among AI researchers. Earlier this year, the company claimed its false-positive rate, wrongly identifying human-written work as AI-generated, was roughly one in 10,000 cases.

Interestingly, older encyclicals written by previous popes reportedly scored as almost entirely human-written when tested with the same system. A transcript of Pope Leo’s public speech was also rated as fully human.

That has only intensified speculation online.

Some observers believe AI may simply have been used as a drafting or editing assistant, much like spellcheck or translation software. Others joked that humanity may have officially entered the era of “AI-generated warnings about AI-generated content”.

Pope asks to slow down on AI

Behind the internet humour, however, the encyclical delivers one of the Vatican’s strongest warnings yet about artificial intelligence.

In the document, Pope Leo calls on world leaders to regulate AI more carefully before commercial competition pushes the technology beyond public control. He argues that ownership of AI systems and data should not remain concentrated in the hands of a few private companies.

The Pope also warned that AI-driven misinformation could worsen political instability and conflict, while the automation race risks harming workers and widening inequality.

Children, he added, require stronger protections from AI-powered platforms and systems.

The encyclical marks the Vatican’s first major document dedicated entirely to AI and arrives as governments worldwide struggle to balance innovation with regulation.