Reflections on Pope Francis's Legacy in AI
Photo by Jeffrey Bruno, released under CC-BY-SA 2.0
Pope Francis died this morning. In this post, I wish to emphasise a lesser-known aspect of his legacy. I’ve recently been reading the statements of religious leaders on AI as part of my research for the upcoming Christians for Impact cause area exploration on AI risk, and Pope Francis stands out among them as perhaps the most articulate on the subject. During the last years of the papacy of Francis, the Catholic Church started to develop its teaching on AI. Pope Francis grasped the transformative potential of AI and called for ethical caution and regulation in its development. Generally, I was surprised by how many themes that Pope Francis discussed were familiar from the AI risk discussions in EA circles.
The engagement of the Catholic Church with AI
The Vatican's interest in AI was sparked around 2016 by the Minerva Dialogues, a series of informal conversations involving people from Silicon Valley. Pope Francis was also approached by ethically-minded business leaders from Europe concerned about emerging issues around AI. Bishop Paul Tighe notes that these early discussions were pivotal in raising awareness within the Vatican about the significance of AI and its potential impact on society.
An important milestone was achieved in 2020 when the Vatican, along with Microsoft, IBM, and other organisations, launched the Rome Call for AI Ethics, outlining principles for ethical AI development. Pope Francis addressed the G7 session on artificial intelligence and devoted his World Peace Day message to AI in 2024. In January 2025, the Dicastery for the Doctrine of the Faith and the Dicastery for Culture and Education issued a doctrinal note called Antiqua et Nova on AI. As far as I’m aware, this is the most extensive official treatment on AI by the Catholic Church. It quotes the Pope’s 2024 addresses frequently.
AI as a “revolution” and bringer of “epochal transformations”
Pope Francis took the transformative potential of AI very seriously. In his address to the G7 leaders, he said that “the advent of artificial intelligence represents a true cognitive-industrial revolution, which will contribute to the creation of a new social system characterised by complex epochal transformations.” As examples, Pope Francis mentioned more equal access to knowledge, giving arduous work to machines and the “exponential advancement of scientific research.”
The mention of exponential scientific progress is especially interesting. One of the things about AI that could bring forward the most change is the possibility of extremely rapid scientific advancement. In a recent publication by Forethought, Fin Moorhouse and Will MacAskill explore this possibility in detail and conclude that the automation of (large parts of) scientific research could enable a century’s worth of progress in a decade. The Pope's perspective here aligns with views commonly held in EA AI safety circles.
In the same address, Pope Francis recognised the uniqueness of AI among technological tools. Simpler tools are under the control of their user, but AI can autonomously adapt to tasks and make goal-oriented choices independent of its user. Without using the word, the Pope is describing agency, which is widely considered a key part of the transformative potential—and the danger—of advanced AI. The Pope also warned against offloading human decision making to AI systems.
However, Pope Francis and other Vatican documents on AI also stress the nature of AI as a tool. Concerns about rogue AI or AI takeover are not mentioned. In line with Pope Francis, the Antiqua et nova doctrinal note emphasises the distinction between human and artificial intelligence and the importance of human uniqueness and dignity.
Calls for regulation
Pope Francis was not an AI doomer, but he very clearly urged caution and supported regulatory frameworks. In his World Peace Day address, the Pope said that AI development cannot be assumed a priori to be beneficial to the future of humanity and that it should be regulated internally by states, and urged the adoption of a binding international treaty to regulate the development and use of AI. I find this quite remarkable, as something like this is precisely what risk-conscious people in AI governance are hoping for instead of the current, essentially unregulated race dynamic. Characteristically, the Pope emphasised that the voices of stakeholders should be taken into account in debates about AI regulation, including “the poor, the powerless and others who often go unheard in global decision-making processes.”
Pope Francis voiced concern about the inequality AI could create. Regarding job loss specifically, he said that there’s a “substantial risk of disproportionate benefit for the few at the price of the impoverishment of many.” Interestingly, Antiqua et nova voices the worry that human labour may lose its economic value when AI becomes more powerful. This echoes some concerns raised in the AI safety community relating to gradual disempowerment scenarios or the so-called intelligence curse.
There’s a stark contrast between the late Pope’s vision of how AI development should be pursued and the current reality. Pope Francis was calling for a saner world where the development of technology with transformative potential is pursued with care, caution and cooperation to ensure it truly benefits humanity.
