The Vatican pushes AI toward a harder test: labor, warfare and human dignity

Pope Leo XIV is putting artificial intelligence where the technology industry rarely likes to see itself: under moral scrutiny from someone who does not need a term sheet. His first encyclical, Magnifica Humanitas, will be presented on May 25, according to The Decoder, and early signals suggest the document will center on human dignity in the age of machine intelligence.

The guest list is the sharper signal. Christopher Olah, an Anthropic co-founder known for work on neural network interpretability, has been invited as a speaker, which gives the event a bridge into the technical safety world rather than leaving it as a purely ecclesiastical statement. That matters because interpretability is one of the few AI safety topics where theology's big language and engineering's small mechanisms can plausibly meet without collapsing into conference fog.

The Vatican is not trying to benchmark a chatbot or bless a model release. It appears to be making a broader claim: that AI systems are no longer just tools for productivity, but institutions of power that shape labor, conflict, and social status. For an audience of 1.4 billion Catholics, that is a distribution channel no AI ethics white paper can casually match.

The hype filter here is useful. Nothing about an encyclical changes a model's training run, a deployment pipeline, or a defense contract by itself. But high-level moral documents can set durable vocabulary, and vocabulary matters when politicians, universities, unions, and companies are still deciding whether AI is a labor story, a safety story, a warfare story, or all three pretending to be a productivity suite.

According to the research brief, Magnifica Humanitas is expected to condemn the use of AI in warfare and address workers' rights. If confirmed in the final text, that would place the Vatican closer to the governance debate than to the usual spectacle of AI optimism, where every system is either an assistant, a copilot, or a mysteriously well-funded destiny. The more interesting comparison is with earlier Catholic teaching on industrialization, which treated technology not as magic but as a reorganization of power.

Olah's presence also gives the presentation a developer-community signal. Anthropic has built much of its public identity around AI safety, and The Decoder's report on the Vatican event suggests the Church wants a conversation that reaches into model behavior, not only ethics vocabulary.

The real signal here is that AI governance is moving out of specialist rooms and into institutions that think in centuries, which is mildly inconvenient for an industry optimized around quarterly demos.