In May 1891, Pope Leo XIII published a document that changed the conversation about work.

Not the work itself. Not the technology reshaping it. The conversation.

Rerum Novarum — On the Condition of Labour — arrived sixty-one years into the Industrial Revolution. The factories had long since transformed the nature of human work. Children were working fourteen-hour shifts in coal mines. Adults were bound to machines in conditions that had no precedent in agricultural or craft economies. The technology was not new. The consequences were not hidden. What was missing was a moral architecture with language precise enough to hold what was happening to people inside it.

The encyclical did not stop the Industrial Revolution. It did not restore pre-industrial conditions of work. What it did was name the thing that had been building for six decades, give it a moral framework that institutions could operate from, and shift the central question from what can this technology do, to what should it be permitted to do to the human experience.

That shift, from capability to permission, is a civilisational threshold.

It happens after the technology is already embedded. After it is already accelerating. After the gap between what the technology is doing and what the human infrastructure can hold becomes visible enough, across enough lives and enough contexts, that it can no longer be deferred.

On Monday 25 May 2026, Pope Leo XIV published Magnifica Humanitas — On the Grandeur of Humanity.

His first encyclical. Addressing artificial intelligence, human dignity, and the conditions of human experience directly.

It was signed on 15 May 2026, exactly 135 years after Rerum Novarum was signed.

That is not a rhetorical coincidence. It is a civilisational signal about what has been crossed.

What the Parallel Actually Tells Us

I am not making a theological argument. I am making a diagnostic one.

When the moral architecture of a civilisation moves to explicitly address a technology, it marks a particular kind of threshold. Not the moment the technology arrives. Not the moment it becomes widespread. The moment when the gap between what the technology is doing and what the human infrastructure can hold forces a naming.

In 1891, that gap had been building for sixty years. Coal, factories, the concentration of industrial power, the conditions of child and adult labour. The technology had moved far faster than the laws, customs, and moral frameworks needed to govern it. Rerum Novarum did not create that gap. It closed the naming distance.

In 2026, the gap is smaller in time but no less significant in kind. AI has been reshaping the nature of work, decision-making, identity, and meaning at a pace that has left governance, law, and moral clarity running behind. The technology has been embedded in hiring systems, customer service, medical diagnosis, content production, financial modelling, and legal analysis, operating on people’s lives before anyone has formally asked whether the conditions are right for it to do so.

What the 135-year parallel tells us is this: civilisations draw this line twice. Once when the technology arrives. Once when the technology reaches the threshold where it is operating on the conditions of human dignity at scale. The first line is about capability. The second is about permission. The second is always the harder one, and it always arrives later than it should.

We are at the second line now.

The question was always coming. The question of what should AI be permitted to do to the human experience. The encyclical is the signal that it has arrived.

What Magnifica Humanitas Names Without the Theology

I have spent more than thirty years preparing organisations for what is arriving, not what is already here. In that time, I have learned to pay close attention to the moments when moral architecture and technical architecture arrive at the same question from different directions. That convergence is always significant, regardless of the institution doing the naming.

Magnifica Humanitas addresses several things directly.

It addresses human dignity as a category that AI systems can threaten or protect depending on how they are designed and deployed. Not theoretically. In practice, through hiring systems that filter without explanation, through content recommendation systems that shape identity and belief, through productivity systems that reduce human beings to performance metrics and treat the deviation from those metrics as a failure to be corrected rather than a human to be understood.

It addresses human agency. The right of people to retain meaningful choice over their own lives, their own work, and the systems that govern them. This is not a demand for a world without AI. It is a demand for a world where AI does not hollow out the conditions that make human choice meaningful, where optimisation does not become the erasure of the person being optimised.

It addresses meaning. The argument that work has never only been about output, and that a future which automates away the conditions that make work feel worthwhile is not progress in any sense that matters to the people living inside it. Efficiency cannot measure meaning. The fact that a system can do something faster and more accurately than a human being does not answer the question of whether the human being still has a role, and therefore a stake, in the outcome.

The presence of Christopher Olah, co-founder of Anthropic, at the presentation of the encyclical is not incidental. It signals that the moral architecture and the technical architecture are now in the same room, asking the same question, using different language to arrive at the same place.

The question: what kind of future are we building, and can people actually live inside it?

The Inhabitable Futures Lens

This is where I want to be specific about what I mean, because the language of Inhabitable Futures is the language I have been developing for exactly this question.

I have a framework I call Inhabitable Futures. It emerged from fifteen years of work as a prison chaplain, sitting with people who had lost the sense that any future was available to them. Not abstractly. Viscerally, in a way that no strategy document or efficiency metric could describe. The opposite of an inhabitable future is not a dangerous future. It is one where trust is absent, agency has been removed, meaning has been hollowed out, and the person is treated as a means rather than an end. I have seen that condition in its most extreme form. I have watched what it does to people.

The framework asks one question: can people actually live, trust, work, and find meaning inside the future being built?

Not is it technically possible. Not is it economically efficient. Is it liveable?

A future is inhabitable when four things are present.

Trust is present. People understand enough of how things work to place reasonable trust in the systems, institutions, and technologies they depend on. They do not need to understand everything. They need to feel that the things they cannot see are not working against them. When an AI system makes a decision about a person’s credit application, their job interview outcome, their medical treatment pathway, or their insurance claim, and that person has no way of understanding why, trust is being withdrawn without consent. The system may be accurate. It may even be correct. But it has failed the trust test, and accuracy without trust is a system people will eventually resist or route around.

Agency is preserved. People retain meaningful choice. Not infinite choice, but enough autonomy that they feel they are living their lives rather than being processed through systems designed by others for purposes that do not include their wellbeing. When AI optimises for engagement over wellbeing, for throughput over judgement, for pattern-matching over individual circumstance, agency is being removed from the people the systems are supposed to serve. The removal is often invisible, because it happens through design choices that feel like convenience rather than constraint.

Meaning is available. Work, relationships, community, and contribution still offer ways to matter. The future has not automated away the things that make human experience feel worthwhile. This is the argument that the efficiency conversation consistently avoids, because efficiency frameworks have no category for meaning. You cannot optimise for it. You cannot measure its absence until the absence has consequences — and by then, the design has already been built.

Dignity is intact. People are treated as ends, not means. The organisations, technologies, and institutions of the future serve human needs rather than extracting human value. This is the language Magnifica Humanitas uses explicitly. Human dignity is not a compliance requirement. It is the standard worth designing toward.

Rerum Novarum used different language. But it was making the same argument across all four dimensions. Trust in labour conditions. Agency for workers. Meaning in work. Dignity against exploitation. The civilisational threshold in 1891 was reached when these four things were being systematically denied by the economic and technological transformation of the era.

The question worth asking seriously in 2026 is whether our AI systems are systematically protecting them.

The Trust Architecture Problem

The research on where leaders draw the line found something that maps directly onto the Inhabitable Futures framework.

Across 120 senior leaders surveyed in Australia, the United States, and beyond, comfort with AI delegation was high for routine, low-stakes tasks. But it collapsed, suddenly, at the point where AI decisions touched identity, livelihood, and reputation. Zero per cent of government and policy executives trusted AI with capital spend decisions. Not low confidence. Not hesitation. Zero.

This is the Trust Cliff. Not a gradual decline in confidence as the stakes rise. A sudden threshold, the point where trust collapses because the human infrastructure needed to hold high-stakes decisions has not been built, and the people in the system sense that absence even when they cannot name it.

The Trust Cliff exists because we have deployed AI faster than we have built the moral and governance architecture that makes high-trust AI deployment possible. We have given people access to AI-assisted systems while withholding the thing those systems require to be trusted: transparency, accountability, explainability, and the consistent evidence that the system is designed with the person’s dignity in mind rather than indifferent to it.

The deeper problem is that organisations have been treating the Trust Cliff as a change management problem. It is not. It is a design problem. You do not overcome it by communicating better about AI adoption. You overcome it by building AI systems that actually pass the four-dimension inhabitability test — systems that preserve trust, protect agency, enable meaning, and respect dignity.

Magnifica Humanitas is, structurally, naming the standard that those systems must meet. Not by restricting AI. By insisting that AI systems be designed to meet the conditions that make trust rational and sustainable.

What the Parallel Teaches About Trust as a Currency

I have spent a long time watching how trust operates as a currency in organisations and markets. The pattern is consistent.

Trust is not primarily built through competence. A system can be accurate and still be distrusted. A leader can be technically capable and still be doubted. Trust is built through reliability, transparency, and the consistent evidence that someone or something is operating with your interests genuinely in frame, not merely as a stated intention.

AI systems, as currently deployed in most organisations, fail this test at exactly the moments when it matters most. A system that performs well on routine tasks but operates without transparency on high-stakes ones is not building trust. It is consuming it. Every time the system makes a consequential decision that the person affected cannot understand, interrogate, or appeal, the trust account is drawn down.

The moral architecture catches up when that draw-down has been running long enough, across enough people and enough contexts, to become structurally visible. In 1891, it was visible in the conditions of industrial labour. The bodies, the hours, the child workers, the absence of any mechanism for workers to have a voice in the conditions of their own work. The draw-down had been running for sixty years.

In 2026, the draw-down is visible in the growing body of research showing that high-stakes AI deployment without transparency and accountability structures consistently erodes the trust of the people it operates on. It is visible in the Trust Cliff data. It is visible in the public scepticism that accompanies every high-profile AI failure involving a real person’s livelihood or dignity.

The encyclical names this, not as a technical problem to solve, but as a design principle to apply. Build AI that people can trust. Build it by designing for dignity, not just for efficiency.

The Ripple Effects from Here

Rerum Novarum did not end child labour on the day it was published.

What it did was shift the moral authority of the question. Before the encyclical, the labour conditions of the Industrial Revolution were a problem some people acknowledged and others disputed, with those in power generally on the side of disputing. After it, they were a problem that the civilisational moral architecture had named in a way that could not be unnamed. The laws, the unions, the safety standards, and the governance frameworks that followed over the next fifty years were the encyclical’s ripple effects, unfolding at the pace that institutions move.

Magnifica Humanitas will follow the same pattern.

In the short term, it will be cited. In parliaments, in boardrooms, in community forums, in academic papers, in policy debates. The language it uses will enter the vocabulary of the AI governance conversation not because everyone agrees with the theology, but because it has given precise moral language to a set of concerns that were previously being expressed imprecisely.

In the medium term, it will shift what is permissible to demand. Employees will cite it when pushing back against AI surveillance in the workplace. Patients will cite it when demanding explanation of AI-assisted diagnoses. Regulators will cite it when designing transparency requirements. Job applicants will cite it when challenging AI-driven hiring decisions. The moral architecture, once named, becomes a resource that people use.

In the longer term, the deeper ripple effect is this: the question of whether a future is inhabitable will move from the periphery of strategic planning to its centre. Not as a compliance requirement. Not as a risk management exercise. As a genuine strategic filter that organisations apply before they build, before they deploy, before they optimise.

The organisations that get ahead of this are the ones that start asking the Inhabitable Futures question now, before the governance frameworks require it, before the Trust Cliff becomes a reputational crisis rather than a design opportunity. The preparation advantage, in foresight terms, belongs to the organisations that read the civilisational signal and act on it while there is still time to act well.

What Organisations Need to Ask Now

The practical question is not what Magnifica Humanitas means for your AI strategy. That is already one level of abstraction too many.

The practical question is this: can the people inside your organisation, and the people your organisation serves, actually live, trust, work, and find meaning in the future you are currently building for them?

That question has four dimensions, and each one deserves a genuine answer rather than a policy document.

On trust: do the people affected by your AI systems understand enough about how they operate to place reasonable trust in them? Not everything. Enough to know that the system is not working against them. If the answer is no, that is a transparency problem. It is not a communications problem — the fix is not better messaging. The fix is building systems that genuinely earn trust through explainability and accountability.

On agency: do people retain meaningful choice over the things that matter to them, or are AI systems optimising away the conditions of choice without the people being optimised noticing? If the answer is uncertain, that is a design problem. Meaningful agency requires that the human being affected by a decision has a genuine stake in, and some form of influence over, that decision. Automation that removes the stake without asking is not a feature. It is a trust withdrawal.

On meaning: do the roles, contributions, and forms of work you are designing toward still give people ways to matter? To exercise judgement, build relationships, develop expertise, recover from failure, and carry responsibility? If the answer is not yet clear, that is a strategy problem. The time to address it is before the roles have been automated, not after the people living in those roles have had their sense of purpose removed.

On dignity: are your AI systems designed to serve the people they affect, or to extract value from them? Is the person at the end of the system treated as an input to optimise or as a human being with a stake in the outcome? If the design does not explicitly protect dignity, it is very likely eroding it, often through choices so small and distributed that no single decision feels significant until the cumulative effect is visible.

These are not soft questions. They are the hardest strategic questions in organisations right now, and most organisations are answering only the first layer: efficiency, cost, capability, competitive position. The second layer, the inhabitable layer, is where the long-term licence to operate is actually determined.

Why This Moment Is Different

I have watched the AI ethics conversation for a decade. For most of that time, it has been a specialist conversation, academics, policymakers, and researchers asking important questions that most organisations were comfortable observing from a distance without fully engaging.

What changes when the moral architecture catches up is that the conversation becomes structurally unavoidable.

This is not about the Catholic Church specifically. What matters is the structural signal: a moral institution that moves slowly and deliberately by design, that has resisted many faster-moving cultural pressures for centuries, has arrived at the same point that the fastest-moving technology companies are currently at. That convergence signals that we are past the point where the AI question is a specialist question or an optional one.

The civilisational institutions are now in the conversation. That means every other institution will be required to respond, not because the Church tells them to, but because the moral language that the Church has given the question will be picked up by every regulator, every activist, every employee representative, and every community member who has been waiting for that language to become available.

The threshold has been crossed. The question has changed.

Not what can AI do.

What should it be permitted to do to the human experience?

Every leader, every organisation, every board needs a genuine answer to that question. Not a principle statement. Not an ethics policy. A genuine, operationalised answer built into the design of every system, every deployment, every decision about what to automate and what to keep human.

The Standard Worth Designing Toward

For thirty years I have been helping organisations prepare for what is arriving, not what is already here. The value of foresight is not prediction. It is preparation.

This week, preparation means something specific.

It means looking at the AI future you are building and asking whether it is one your people can actually live inside. Not whether it is efficient. Not whether it is competitive. Whether it is liveable, trustworthy, meaningful, and respectful of the dignity of the people who will have to inhabit it.

The moral architecture just named that question out loud, at civilisational scale, in a way that 135 years of history tells us marks a turning point.

A future people can trust. A future that preserves their agency. A future where meaning is still available. A future where their dignity is intact.

That is the Inhabitable Futures standard.

It is the standard worth designing toward.

Prepare now.

Choose Forward.