The first AI encyclical

On May 25 the Vatican released Magnifica Humanitas, the first papal encyclical devoted to artificial intelligence. An encyclical is the Church's top-rank teaching document, and the choice of genre is the news: the Vatican had spoken on AI before, but only through advisory notes and conference messages. This commits the Church's full magisterial authority to a set of demands.

It is "not permissible to entrust lethal or otherwise irreversible decisions to artificial systems."

The demands are concrete. Governments, not the firms, must hold the 'regulatory tools' over the AI boom. Workers displaced by automation are owed protection — Leo signed the document on the 135th anniversary of Rerum Novarum, his predecessor's 1891 charter on industrial labor, casting AI as the same kind of shock to ordinary work. And on autonomy the line is flat: it is 'not permissible to entrust lethal or otherwise irreversible decisions to artificial systems.' The text also refuses AI any personhood — these systems 'do not undergo experiences, do not possess a body... Nor do they have a moral conscience.'

The lobbying around it inverts what you'd guess. Executives from Google, Amazon, and Meta met the Pope and then spent hours with a Vatican communications official at the French embassy in Rome. Anthropic took the opposite tack — it had already folded the Vatican's ethical vision into the written 'constitution' that shapes how its Claude models are trained, and sent a co-founder, a 33-year-old atheist, to speak beside the Pope at the launch. A frontier lab is encoding Catholic-derived ethics into a model while the other giants worked the back channels.

For 1.4 billion Catholics and for policymakers looking for a high-authority reference, the encyclical hands them a fixed position on the questions still open at the UN and in every legislature: who is liable, what machines may decide, and whether a chatbot can ever be a someone.