A more moral AI?
Pope Leo is right to worry about Artificial Intelligence. But whose morality should guide it?
The Holy See has issued an encyclical about safeguarding the human person in the time of Artificial Intelligence. You can read it here.
It describes in a broadly accurate way how modern AI works, and how it differs from older rules-based technologies.
It raises thoughtful concerns about the potential impact of AI, which many atheists and secularists would share.
It builds those concerns on Catholic theological arguments that are not based on publicly testable evidence.
Its concerns about AI having opaque and unaccountable power could also be a mirror for the Church’s own governance.
It’s an odd document to read as a secularist fascinated by technology. I get the same feeling, on an intellectual level, as I do emotionally when I go to a funeral mass. There, I hear someone speak meaningfully and knowledgeably about a dead friend. Then the priest says something about him being up in heaven looking down on us.
It’s like two very different conversations overlaid on each other. Religious and secular arguments often share the same sentence, with different unspoken rules about how you should justify the various claims. So let’s try to disentangle it.
How Artificial Intelligence works
The encyclical impressively highlights key aspects of AI, including:
“98. All of us, including those who design them, possess only a limited understanding of their actual functioning. Indeed, current AI systems are more “cultivated” than “built,” for developers do not directly design every detail, but instead create a framework within which the intelligence “grows.” As a result, fundamental scientific aspects — such as the internal representations and computational processes of these systems — remain, at present, unknown.”
“99. They often surpass human intelligence in speed and computational capacity, offering tangible benefits across many fields. Yet this power remains entirely tied to data processing… Even when these tools are described as capable of “learning,” their way of doing so is different from that of a human person… Rather, it is a form of statistical adaptation based on data and feedback.”
These are broadly accurate if incomplete ways of describing AI. ‘More cultivated than built’ captures the essence of how modern AI differs from older rules-based technologies.
I find chess programs a useful analogy in this respect. Older programs relied heavily on human-designed rules, brute-force search, expert tuning, and opening books and endgames based partly on master games. More recent AI-style systems such as AlphaZero can start with little more than the rules of chess and improve by repeatedly playing themselves.
The OECD describes AI like this:
“An AI system is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment.”
The OECD version is more precise, but the encyclical captures the broad shift from explicit programming to systems that infer patterns from inputs and generate outputs. In many modern systems, developers do not hand-code every insight. They create the conditions in which the system can learn effective patterns.
Concerns about the social impact of AI
The encyclical highlights thoughtful concerns about the social impact of AI, including:
“101. We see that it is now embedded in decision-making processes across many sectors and at multiple levels: in communication, management and control.”
“132. Disinformation did not begin with AI, yet today it finds a powerful amplifier in AI. The ability to manipulate content, images and videos exposes people to biased or misleading perspectives.”
“102. When AI systems present themselves as neutral and objective, they end up reflecting and reinforcing the stereotypes or ideological bias of their designers and developers.”
“132. Only the shared pursuit of the veracity of facts, perceived as a common good, can provide a solid foundation for just communication.”
“101. Current AI systems require enormous amounts of energy and water, significantly influencing carbon dioxide emissions, and place heavy demands on natural resources.”
“197. the growing ease with which autonomous weapons systems can be deployed makes war more “feasible” and less subject to human control.”
Many secularists share most or all of these concerns. I don’t have the same level of concern as the encyclical about some specific claims. Some are phrased quite generally, as if all types of AI are the same and cause the same risks. But they are all thoughtful concerns that society should address.
The encyclical is built on Catholic theology
Because its broad analysis of AI is good, it is easy to forget that this encyclical is at heart a theological argument. Like all encyclicals, its purpose is to explain, strengthen, and evangelise the Catholic view of the world including its untestable claims about a supernatural reality. As the introduction puts it:
“1. Humanity, created by God in all its grandeur, is today facing a pivotal choice: either to construct a new Tower of Babel or to build the city in which God and humanity dwell together.”
“6. For this reason it is necessary to begin a shared discernment process for identifying the spiritual and cultural roots of ongoing transformations.”
“7. In order to answer these questions and discern how to navigate responsibly the era of AI, I would like to bring to mind two scenes from the Bible: the construction of the Tower of Babel (cf. Gen 11:1-9) and the rebuilding of the walls of Jerusalem (cf. Neh 2–6).”
“9. Therefore, the primary choice is not between a “yes” or “no” to technology, but rather between constructing Babel or rebuilding Jerusalem; between a power that claims to dominate the heavens and a people who work together in the presence of God to rebuild the walls of fraternal coexistence.”
“11. Building a city founded on the common good implies, first and foremost, building on a firm relationship with God.”
“14. Finally, building for the common good requires an evangelical language.”
The encyclical repeatedly mixes this type of evangelical theological language with its technical analysis of AI and its social impacts. Wherever it refers to concepts like the dignity of the human person, the universal destination of goods, or care for our common home and peace, it means what these concepts mean in Catholic theology, not secular philosophy or human rights law.
This complicates any attempts, which would be worthwhile if possible, for secularists and the Catholic Church to build coalitions around shared concerns about the practical impacts of AI. Because once we start to address how, rather than whether, to address the concerns, the difference between what we each mean by these concepts will come into play.
The encyclical is a mirror for the Catholic Church
The encyclical makes several perceptive analyses of the impact of AI on how society tries to identify what is true, and to navigate moral disagreements. The Holy See could also apply these analyses to its own church governance.
“95. These entities effectively set the conditions for access, determine the rules of visibility and shape the very possibilities for participation. When such power is concentrated in the hands of a few, it tends to become opaque and evade public oversight, increasing the risk of distorted forms of development that give rise to new dependencies, exclusions, manipulations and inequalities.”
“105. In many cases, however, the internal processes leading to a result remain opaque, making it harder to assign responsibility and correct errors. This is where accountability becomes crucial: the possibility of identifying who must “account” for decisions, justify them, monitor them, and, when necessary, challenge them and remedy any harm caused.”
“107. [We must insist on] openly discussing the ethical frameworks involved and subjecting them to shared standards of social justice. Otherwise, those who control AI will impose their own moral vision, which will become the invisible infrastructure of these systems. A more moral AI is not enough if that morality is determined by a few.”
“132. The quality of public communication depends directly on social trust and, in turn, shapes it. At the same time, truthful information does not arise from centralised or automated control. In public discourse, the truth of facts has a rational dimension, as it requires verification, cross-checking of sources and responsible argumentation… Only the shared pursuit of the veracity of facts, perceived as a common good, can provide a solid foundation for just communication.”
“133. Those who command powerful technological and economic resources, along with substantial human capital for intervention, possess significant capabilities for influencing cultural change. Ultimately, they can influence a significant number of people concerning the truth about humanity, the world, the meaning of existence, the family and even God. This is pure power detached from truth, which subtly or overtly imposes what it wishes others to accept as true.”
In fairness, the encyclical does say that Christian communities, too, are called to commit themselves to transparency in communication and to the honest pursuit of facts. But it does not apply that standard to its own doctrinal beliefs about truth and morality.
This raises an obvious question. If the Holy See believes we should openly discuss ethical frameworks that shape human lives, and subject those ethical frameworks to shared standards of justice, why should the Catholic Church’s own moral teachings be exempt from the same scrutiny?
Conclusion
The Holy See has produced an encyclical that describes in a broadly accurate way how AI works, and raises thoughtful concerns about its impact on society. Many atheists and secularists share these concerns, but do not share the underlying theological arguments that underpin them.
This encyclical includes a perceptive analysis of the relationship between power, accountability, truth, and justice. The Holy See could use that analysis not only as a warning about AI, but also as a mirror for its own church governance. That would strengthen its criticism of opaque and unaccountable AI power.
Atheists and secularists should not just dismiss this document. Many of its practical recommendations overlap with our concerns about human rights, democratic accountability, transparency, and other risks.
We should find ways to cooperate where our conclusions overlap, while being honest that we reach those conclusions from very different starting points.