Vatican's AI Encyclical Appears Partly AI-Written, Analysis Shows

An investigation into 'Magnifica Humanitas', the recent encyclical by Pope Leo XIV, reveals that significant portions appear to be AI-generated. The document, which emphasizes human dignity and warns against AI replacing human relationships, was analyzed using Pangram, an AI detector known for low false-positive rates. The Italian text—likely the original—was tested section by section. Results show Chapter 1 had the highest AI-flagged content at 62%, followed by the Introduction and Conclusion at 43% each, Chapter 3 at 41%, and Chapter 2 at 34%. Even the lowest chapters (4 and 5) had 24% and 18% flagged, respectively.

The findings raise questions about the Church’s credibility when its own guidance on AI ethics may itself rely on the technology it warns against. The author notes that paragraphs flagged as AI-written are almost certainly so, though some unflagged sections may also be AI-assisted. This incident highlights the growing challenge of detecting AI-generated text in official documents and the ironic potential for institutions to undermine their own messages by using the very tools they critique.