Pope Leo XIV released Magnifica Humanitas, a 40,000-word encyclical written in Latin that condemns current AI practices as threats to human dignity. The document argues that large language models promote dehumanization and calls for concrete policy changes to safeguard personhood. Its arrival marks the first formal papal intervention in AI governance.

Tech firms responded with a mix of skepticism and vague reassurances, while a handful of employment lawyers began probing whether the papal stance could support a Catholic exemption from mandatory AI tools at work. Simultaneously, a consortium of religiously affiliated universities published a study claiming that LLMs rarely produce overtly religious answers, sparking debate over the role of faith‑based perspectives in model training.

Critics warn that the study’s narrow definition of religion—equating it with fundamentalist Christianity—mirrors LLMs’ tendency to generate answers from limited data sets, inviting pressure to embed specific doctrinal content for “fairness.” The Vatican’s warning and the legal chatter illustrate how AI’s opacity is intersecting with religious authority, forcing regulators to confront competing claims today. Policymakers must choose between secular utility and cultural reverence.