Beyond What Is Permitted

The Global Compact on Education in the age of artificial intelligence. Notes after the VII Uniservitate Global Symposium in Castel Gandolfo.

This week I took part in the VII Uniservitate Global Symposium in Castel Gandolfo, a gathering of universities that practise service-learning, i.e. a way of teaching in which students learn by putting what they know at the service of a community. I had been invited to speak in a plenary session on artificial intelligence, around a simple question: what do we need to create a more human digital world? What struck me most, however, were the students themselves, moved by deep ideals and leading projects for their own schools and communities towards a common good.

The answer I tried to give rests on a distinction between what is permitted and what is right. Our societies are responding to AI with regulation (in Europe, the AI Act) and codes of conduct, and our universities are responding with AI literacy, i.e. with courses on how these systems work, how to use them and where they fail. Both responses are necessary, and we will need more of them. However, both work on the first side of the distinction: they tell us what is allowed and how the tool works. The question of what is right, here and now, for the people involved, belongs to conscience, i.e. to our capacity for discernment, and conscience is formed neither by a rule nor by a course on how to use a machine. Leo XIV expressed this clearly in Magnifica Humanitas (§140): every technology educates those who use it, so that educating people to use AI means educating them to decide when, and for what, not to use it.

The Global Compact on Education, which was present throughout the symposium, was proposed by Pope Francis in 2019 and relaunched in October 2020. Its starting point is an African proverb he often quoted, according to which it takes a whole village to educate a child. The Compact asks for an “educational village” in which families, schools, universities, cultures and religions take on together a task that none of them can carry alone, and its commitments begin with placing the person at the centre and listening to the voices of children and young people. Read in the age of AI, this intuition acquires a precise meaning. Indeed, AI literacy can be entrusted to a single course, while the formation of conscience can only be the work of a village.

Within this village, the religious and spiritual traditions can play a specific role. For millennia they have cultivated the interior life of the person, i.e. the place where we decide not only what to do but who we want to be, and they have developed practices for it (the examination of conscience, silence, discernment, the making and keeping of promises). This is precisely the space that neither law nor technology can reach, and in 2021 Francis gathered representatives of different religious traditions around the Compact, recalling that religions have always had a close relationship with education. Their contribution is a possibility, and it depends on their capacity to understand the present, i.e. to bring their experience of the interior life into dialogue with the questions that AI is actually raising. Families, schools, universities and many secular traditions of moral education form conscience as well, and the Compact places religions inside the village, alongside them. When religions take up this task, they can offer a long memory of how a person is formed, together with the capacity to speak to communities that no regulation will ever reach, as many of them have already done in front of the ecological crisis.

This is also why the symposium struck me. Service-learning is, in a sense, a practice of the educational village: students learn in front of real persons and real needs, and in doing so they learn to answer for what they know. Conscience is formed in this way, through discernment exercised together in front of a concrete and difficult decision, and the projects that the students presented in Castel Gandolfo showed it far better than any lecture on ethics could. Should I use AI for this task? Is it right for the people it will affect? Questions of this kind have no answer in a textbook, and they are learned by facing them again and again, with others.

It is in this spirit that, with colleagues in several countries, I am working on an instrument of this kind: an oath for the age of artificial intelligence, i.e. a short set of shared principles that a person can assume publicly, before others who can ask them to honour it. A code is complied with, whereas an oath is assumed, and for this reason it does not regulate behaviour from outside but forms the person who pronounces it. Indeed, a machine can produce the sentence “I promise”, but it cannot promise, and this is why I believe that an oath belongs to the work of the village rather than to that of the regulator.

At the end of my talk I left the students with a question: when did you last decide not to use AI because it was not right, and with whom did you take that decision? I think it is also a question for our institutions, and for the religious traditions that have accepted to be part of the educational village. Answering it requires a community in which such decisions can be discussed and shared, and this is perhaps what the Compact had in mind when it spoke of a village.

Cover: based on the graphic design of the VII Uniservitate Global Symposium, “Designing Maps of Hope through Service-Learning”, Castel Gandolfo, October 2026.