When the papal encyclical “Rerum Novarum” by Leo XIII was released in 1891, it was a milestone not only in the foundation of Christian Social thought at the height of the Industrial Revolution. It was much more than an answer to the specific grievances of the European working class. It was a thorough reflection on the human being and humanity in times of unprecedented technological, economic and social transformation.

His successor, Leo XIV, has now taken up this tradition of rethinking human existence and the essence of a fair society at the doorstep of an era which might undermine the very essence and dignity of humankind. As his latest Encyclica “MAGNIFICA HUMANITAS” correctly puts it: the central challenge of AI is not simply what machines can do or whom they can replace, as happened during the first industrial revolution. The challenge is increasingly what algorithmic automation may do to human agency itself, how decisions are made, how information circulates and is shared, and how power is exercised.

AI algorithms increasingly shape what people see, read, believe and prioritise. At the same time, AI systems depend on vast concentrations of data, infrastructure and computational power, all controlled by a relatively small number of private and state actors. Their influence already exceeds that of many states, as the encyclical rightly warns.

The debate on AI is often framed around efficiency: how to optimise processes, accelerate decision-making, or reduce costs. Yet the Christian social tradition has always insisted that the human person and societies cannot be “measured” solely by efficiency. The point of reference is whether people remain capable of responsibility and whether spaces for participation, meaningful social interaction and responsibility continue to exist.

This is where the principle of subsidiarity, a cornerstone of Christian social thought, becomes especially relevant once again. This search for human-centred social organisations has profoundly shaped Christian-democratic thought in Europe. Subsidiarity is not simply about decentralisation or limiting the (central) state. The principle assumes that persons, families, communities and intermediary institutions should be sufficiently equipped and capable of acting responsibly within society. The goal is not fragmentation, but participation. A healthy society depends on active citizens and living communities rather than passive dependence on distant systems of control.

The encyclical strongly reaffirms subsidiarity as a way of protecting the freedom, creativity and responsibility of persons and communities in the face of and against growing concentrations of power. Artificial intelligence now challenges this principle in ways one could never have foreseen.

At present, AI systems tend towards centralisation. They require enormous investment, large volumes of training data, highly concentrated infrastructure, and large-scale behavioural modelling. The more powerful these systems become, the greater the temptation to transfer elementary human capabilities, such as judgement, coordination and even social trust, from persons and institutions to algorithms.

The societal consequences are dramatic, and the encyclical addresses them to a large extent.

First, AI might weaken human judgment itself. Those systems increasingly recommend what we should read, watch, purchase, or believe. In core institutions of society, administration, education, and even healthcare, algorithms increasingly influence decisions that were previously, and to a large part, made by human beings. A society in which citizens gradually lose the habit and the ability of self-determined judgment undercuts the very foundations of democratic life. A purely technocratic society, as envisaged by some and governed through optimisation and behavioural prediction, may remain efficient (still to be proved) – but definitely less human.

Second, AI could bypass intermediary institutions that traditionally shape civic responsibility and social cohesion: families, schools, local communities, associations, churches, and local media. Until now, they have provided the social spaces in which judgment and participation of (young) people are formed and trained. Digital platforms are increasingly establishing direct algorithmic relationships with individuals. This weakens those mediating and deliberative structures that sustain democratic societies.

These developments are of greatest concern to Christian democracy. As “Rerum Novarum” had already pointed out 135 years ago, Christian-democratic thought must always resist both excessive state centralisation and excessive market concentration. In a very clairvoyant way, it anticipated the threats to a pluralistic social order that undermine strong intermediary institutions, civic participation, and shared responsibility. AI might dwarf the 20th-century attempts to achieve all-encompassing command and control over citizens.

Third, traditional political power is at least formally visible. This creates a dangerous asymmetry. The more society depends on algorithmic systems, the greater the risk that citizens become passive users rather than active participants in democratic life.

The Encyclical does not oppose technological development. Magnifica Humanitas clearly acknowledges the enormous positive potential of technology for education, healthcare, communication and human flourishing. But it insists, in line with a long tradition of Christian social thinking, that technology must remain oriented and ordered toward the human person and the common global good.

The implications for Christian-democratic policy are clear:

• A Christian-democratic approach to AI should therefore focus not only on innovation or (de)regulation, but also on preserving human agency within technological societies.
• Human judgement must remain central in areas affecting dignity, responsibility and democratic participation; it cannot be delegated.
• Digital governance must ensure transparency and accountability so that power remains visible and contestable.
• Intermediary institutions must be strengthened rather than bypassed by technological systems.
• And, as convenient as it often might be in first hindsight, societies must resist the temptation to reduce persons to data points, consumers or objects of behavioural management.

The real AI debate, as the Pope reminds us, is therefore not simply about technology. It is about the kind of civilisation Europe wants to be. Will democratic societies remain communities of responsible persons and living institutions, or will they gradually evolve into systems driven primarily by technological optimisation and concentrated digital power? In the age of artificial intelligence, the core fight is about defending the human person itself.