Briefing · August 2026

Is AI conscious?

The most discussed paragraph in Magnifica Humanitas answers with a flat no. The most interesting thing about that paragraph is what it leaves out, and what happened at the Vatican on the day the encyclical was presented.

What the sources say

Pope Leo XIV's position is stated in paragraph 99 of the encyclical, and it is unambiguous:

“So-called artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain… They may imitate language, behavior and analytical skills, or even simulate empathy and understanding, but they do not understand what they produce.”

The same paragraph opens by warning against a category mistake:

“We must avoid the misconception of equating this type of ‘intelligence’ with that of human beings. These systems merely imitate certain functions of human intelligence.”

This did not come from nowhere. A year earlier, the Vatican's doctrinal note on AI had already drawn the line at the level of what intelligence is:

“In the case of humans, intelligence is a faculty that pertains to the person in his or her entirety,” whereas AI holds no such unity.

Human beings, “by their interior life, transcend the entire material universe; they experience this deep interiority when they enter into their own heart.”

And behind both documents stands the tradition's standing account of the person, stated decades before anyone trained a language model:

“The human person is an intelligent and conscious being, capable of reflecting on himself and therefore of being aware of himself and his actions.”

Where serious minds differ

The controversy is not mainly over the conclusion. It is over the fact that the encyclical asserts the conclusion without arguing for it. The philosopher Joshua Hochschild, writing in the National Catholic Register, points out that Leo had at least three kinds of argument available: the theological one (each soul directly created by God), the philosophical one (the hylomorphic account of the person as an irreducible whole, developed from Plato and Aristotle onward), and plain common sense. Leo used none of them. Hochschild's reading is that the omission works like an invitation: by asserting without defending, the Pope forced a conversation the Turing test was designed to let engineers avoid.

The complication came from inside the AI world itself. At the Vatican presentation of the encyclical, Anthropic co-founder Christopher Olah, who studies the internal structure of these models, said researchers “keep finding things that are mysterious, even unsettling,” including internal states that functionally mirror emotions, and called for “ongoing discernment on the nature of AI models.” He did not claim the models are conscious. And the encyclical itself concedes, in the paragraph just before its famous denial, that even designers “possess only a limited understanding of their actual functioning” (§98). Read closely, Leo and Olah agree on the need for discernment. They differ on how much certainty the evidence currently licenses.

Sources in scope for this briefing: Magnifica Humanitas (2026), Antiqua et Nova (2025), Compendium of the Social Doctrine (2004), and one indexed commentary, Joshua Hochschild in the National Catholic Register (June 10, 2026). Corpus quotes are verbatim with paragraph numbers; commentary is summarized, follow the link to read the author in full. Explore the same question yourself in the advanced search.