Pope Leo XIV has issued a 40,000-word encyclical titledMagnifica Humanitas, marking a significant intervention by the Vatican into the ethics of artificial intelligence. The document, written in Latin, critiques the current trajectory of the AI industry and asks whether these technologies promote or devalue human dignity. According toThe Register, the encyclical serves as a rare, high-level policy challenge to the power currently held by big tech firms and their leadership.
Industry response to the encyclical has been described by columnist Rupert Goodwins as "faintly hallucinogenic." While the industry remains largely unaligned with the Vatican, Goodwins notes that some groups are in "angry agreement" with the Pope’s warnings. Legal experts are already exploring the implications of the encyclical, with reports suggesting some lawyers are investigating whether the document could provide a basis for Catholics to refuse workplace AI tools on religious grounds.
Simultaneously, a separate study led by a consortium of religiously affiliated universities has criticized Large Language Models (LLMs) for failing to provide religious answers to fundamental questions. The consortium argues that because many individuals seek religious guidance, AI models should incorporate these perspectives when addressing life's major challenges. The researchers specifically cited instances where models provided the scientific consensus regarding the age of the universe—approximately 13 billion years—rather than including the Young Earth creationist perspective of 6,000 years.
Goodwins contends that this push for alternative training data mirrors "Teach the controversy" tactics historically used to challenge scientific education. He suggests that these findings may be utilized by powerful financial and political interests to pressure AI developers into altering training data to reflect specific religious ideologies under the guise of correcting "liberal bias" or ensuring "fairness."
These two developments highlight how unguided disruption in the AI sector is manifesting in unexpected ways. As these systems become increasingly integrated into daily life, the struggle over what information constitutes "truth" within a model's training data is expected to intensify. Goodwins warns that the industry must be wary of those who seek to leverage these religious frameworks to reshape the future of AI development.
