Pope Leo XIV's first encyclical may have been partially written by artificial intelligence. A detailed analysis by researcher Linch Zhang, shared on the LessWrong forum, used the AI detection tool Pangram to examine the document called Magnifica Humanitas. The results suggest portions of the encyclical show a 40 to 100 percent likelihood of being AI-generated. Zhang specifically noted an unusual frequency of the word "genuinely" — a term commonly associated with writing produced by Anthropic's Claude AI — that doesn't appear at similar levels in the pope's previous works. When The Verge independently tested roughly 2,000 words from the document through Pangram, the tool estimated 46 percent was AI-written. A separate analysis of the encyclical's first chapter showed an even higher rate, with 62 percent flagged as AI-generated. However, the results weren't uniform across the entire document — some sections registered as essentially 0 percent AI-written, and Pangram rated the first 20 paragraphs of the last four encyclicals as 100 percent human-written with high confidence. A transcript of a Pope Leo speech was similarly rated as entirely human. The Vatican's office for communications didn't respond to requests for comment. While AI detection tools aren't infallible and different systems often produce varying results, Pangram has earned respect among AI researchers. The company reported in March 2025 that it estimates its false positive rate — incorrectly flagging human-written content as AI-generated — at approximately 1 in 10,000. Encyclicals are formal letters issued by the pope to address significant moral and social challenges of the era. This marks the first encyclical from Pope Leo XIV, with the most recent previous one authored by Pope Francis in October 2024. Notably, Magnifica Humanitas is the first papal encyclical to center specifically on artificial intelligence and its broad impacts on humanity — making the potential use of AI in drafting the document somewhat ironic given its critical subject matter.