While governments debate regulation and corporations race toward the next breakthrough, Pope Leo XIV’s encyclical Magnifica Humanitas arrives.

It asks something more uncomfortable: not what artificial intelligence can do, but whom it will do it to.

Most people, if they are honest, have not thought seriously about that distinction. They experience AI as a convenience — a search that anticipates, a translation that arrives instantly, and a recommendation that somehow knows.

What they do not see, because it has been designed not to be seen, is the vast asymmetry underneath.

Someone built these systems. Someone owns them. Someone decided whose knowledge would train them and whose would not.

And those decisions, made quietly in boardrooms and research labs mostly concentrated in a handful of cities, are quietly shaping what the rest of the world will be allowed to know.

A pattern history repeats

There is a reason the phrase ‘AI colonialism’ makes people uneasy. It is not because the comparison is unfair. It is because it is uncomfortably accurate.

Colonial systems did not announce themselves as exploitative. They arrived with the language of progress, civilisation, and mutual benefit.

What they actually did was extract resources, labour, land, and knowledge from societies that lacked the power to refuse, while concentrating the returns elsewhere.

The affected communities were not absent from the system. They were essential to it. They simply did not control it, and they did not meaningfully benefit from it.

There is a reason the phrase ‘
AI colonialism’
makes people uneasy.
It is not because the comparison is unfair.
It is because
it is uncomfortably accurate.

The data economy has reproduced this structure with remarkable fidelity. Communities across the developing world generate enormous quantities of information daily — their languages, habits, relationships, and cultural expressions continuously feeding systems they did not design and will not govern.

Magnifica Humanitas does not use the word ‘colonialism’. It does not need to.

Pope Leo’s warning against new forms of domination emerging through concentrated technological power describes the same dynamic with greater precision and considerably more moral seriousness.

What dignity requires

The encyclical’s most grounding conviction is also its most countercultural in the present moment: human beings are not resources.

This sounds self-evident. It is not.

The entire architecture of the digital economy rests on a contrary assumption — that personal data, generated by billions of people going about their lives, is a raw material available for collection, processing, and conversion into profit by whoever has the infrastructure to do so.

The person at the centre of that exchange is rarely consulted. Almost never compensated. Structurally invisible.

Pope Leo insists on a different account. Human dignity, in the Catholic tradition he draws on, is not conferred by productivity or market value.

It is inherent, which means it cannot be legitimately overridden by commercial interest or geopolitical strategy.

That is not a sentiment. It is a boundary. And it has direct, uncomfortable implications for systems that currently treat the data of the world’s poorest communities as freely available raw material.

Knowledge being lost

Running quietly alongside the economic injustice is something that may prove harder to recover: the erasure of ways of knowing that were never fed into the machine.

Artificial intelligence learns from what it is given. When training data skews heavily toward certain languages, institutions, and cultural frameworks, the resulting systems carry those skews forward — not as acknowledged bias but as assumed universality.

The model does not know it is partial. It presents its conclusions as simply correct.

Indigenous knowledge systems, oral traditions, minority languages, and local ecological understanding accumulated across generations — none of these are marginal curiosities.

They represent genuinely different ways of interpreting the world, different questions worth asking, and different answers that dominant systems have not yet thought to look for.

When they are absent from the systems now being built, they are not merely under-represented. They are being written out of the future.

The future of human knowledge
has never truly belonged to the powerful,
even when they acted as though it did.

An inheritance for all

Magnifica Humanitas speaks of humanity’s collective wisdom as a shared inheritance. That word — inheritance — carries weight.

An inheritance is not something a minority administers on behalf of everyone else while quietly deciding what gets preserved and what gets discarded. It belongs to all, and the decisions made about it should reflect that belonging.

Beyond competition

The international race for AI dominance has a logic that is easy to follow. Nations that fall behind risk dependence. Nations that lead gain influence.

The strategic incentives are real and should not be dismissed.

But Magnifica Humanitas challenges the assumption beneath the logic — that this kind of competition is either inevitable or, more importantly, good.

Solidarity over rivalry

The Pope proposes solidarity as an organising principle, and he does so without sentimentality. Solidarity does not mean the absence of national interest.

It means recognising that some challenges are too consequential, and some opportunities too important, to be managed through rivalry alone.

An equitable AI future would look concretely different from the one taking shape. Developing nations would have genuine access not just to AI products but also to its governance.

Communities would retain meaningful ownership over the data they generate. International frameworks would prioritise knowledge-sharing over knowledge-hoarding.

None of this is technically out of reach. It is politically difficult, which is a different problem, with different and more tractable solutions.

The only measure

Pope Leo does not offer a policy framework, and that is not a weakness. What he offers is something prior to policy: a moral standard against which whatever frameworks emerge can be honestly judged.

By that standard, the question is not whether AI is powerful or innovative or economically transformative. It clearly is all of those things.

The question is whether that power is being directed toward the flourishing of every person — including those with no leverage, no infrastructure, and no seat at the table where the decisions are made — or whether it is reconstructing old hierarchies in a new and more durable form.

The future of human knowledge has never truly belonged to the powerful, even when they acted as though it did.

It belongs to everyone who has ever tried to understand the world and pass that understanding on. The digital age does not change that. It only makes the stakes of forgetting it considerably higher.

• John Singarayar SVD holds a doctorate in Anthropology. He is an author of several books and regularly contributes to academic conferences and publications focusing on sociology, anthropology, tribal studies, spirituality, and mission.