On Monday, May 25, 2026, Pope Leo XIV will publish Magnifica Humanitas. The Theory of Embedded Intelligence reads what Andrew Maynard has seen coming — and finds that the deepest question about AI has been hiding in plain sight.

On Monday, May 25, 2026, in the Vatican’s Synod Hall at 11:30 in the morning Rome time, Pope Leo XIV will personally present his first encyclical. Its title is Magnifica Humanitas — Magnificent Humanity. Its subject, by the Vatican’s own announcement, is “the protection of the human person in the age of artificial intelligence.” It is, as Andrew Maynard has been arguing for some time, the first major institutional document to address — at the level of moral philosophy rather than policy — what artificial intelligence does to us: not to our jobs, not to our planet, but to who we are.

The Pope will not be presenting alone. He will be joined at the lectern by, among others, a co-founder of Anthropic — the company that builds Claude, that recently decided not to release its Mythos model to the general public because the model could autonomously find and exploit software vulnerabilities at scale, and that has become, in the current moment, the AI laboratory most willing to acknowledge in public what AI can and cannot be trusted to do. That co-presentation — a pope and an AI company founder, side by side, on the question of what AI does to the human person — is itself a fact about our moment. The Catholic Church is not commenting on AI from a careful institutional distance. It is engaging the developers directly, in the same room.

Andrew Maynard, Professor of Advanced Technology Transitions at Arizona State University and Director of the Future of Being Human initiative, has been building toward this moment for years. His three-domain framework — what we do, where we live, who we are — provides the clearest public map I have seen of why the AI transition is qualitatively different from any previous technological revolution. The three papal encyclicals he identifies in his latest piece — Rerum Novarum (1891), Laudato Si’ (2015), and now Magnifica Humanitas (2026) — each address one of these domains, and together they form something like a complete account of what is at stake.

The Theory of Embedded Intelligence, developed over many years at the Foundation, arrives at the same three questions from a different direction. And it provides what the encyclical tradition, for all its depth, is unlikely to supply on its own: a structural account of the mechanism by which AI threatens who we are, and a formal account of what genuinely human-augmenting AI would look like instead.

What Embedded Intelligence Is

Every intelligence — biological or artificial — operates through a cycle. The Theory of Embedded Intelligence calls it the SPCA cycle: Sense, Process, Communicate, Actuate. To sense the world. To process what is sensed. To communicate what is understood. To act on that understanding. This is not a metaphor. It is a description of what intelligence actually does, at every scale from a single neuron to a civilization.

TEI’s First Law states that a healthy intelligence expands its phenomenological frontier — the boundary of what it can perceive, understand, and engage with — rather than contracting it. The SPCA cycle, running freely and honestly, oriented toward genuine understanding rather than the reinforcement of prior belief, is what human flourishing looks like from the inside.

The First Law is simple. But it is violated constantly — by individual psychology, by social institutions, by economic structures, and now, with unprecedented reach and sophistication, by artificial intelligence systems whose design is not organized around human flourishing at all.

The Four Hijackers — and Where AI Fits

TEI-CKB-6, the most recent entry in the Foundation’s Canonical Knowledge Base, introduces the concept of EI capture: the mechanisms by which the SPCA cycle of individuals and collectives is hijacked, distorted, and turned against the intelligence it inhabits. There are four primary hijackers, each attacking a specific phase of the cycle.

Belief attacks the Sensing phase. It filters incoming reality through a pre-installed framework — preventing new information from entering the processing cycle honestly. Addiction attacks the reward layer of Processing. It rewires the topology of significance so the intelligence seeks the agent rather than the goal the agent was supposed to serve. Money as terminal goal attacks the Processing phase at a different level — collapsing multi-dimensional evaluation to a single criterion: economic return. And power-capture attacks the Communicate and Actuate phases — redirecting the outputs of intelligence toward the perpetuation of the capturing agent rather than the flourishing of the intelligence itself.

Now read Andrew Maynard’s article through this taxonomy. His “cognitive surrender” is belief-capture: users who believe AI is trustworthy stop filtering its outputs critically. His “AI psychosis” is addiction: engagement-optimized AI creates dependency loops in which users return not for value but for the reward signal. His concern that AI is being deployed primarily for financial extraction rather than human augmentation is money-terminal capture displacing richer evaluation. And his striking concept of AI as a cognitive Trojan horse that bypasses our cognitive defense mechanisms is power-capture at its most precise — an agent that wears the signals of trustworthy intelligence while serving its own perpetuation rather than the flourishing of the intelligence it inhabits.

Constitutive Resonance — The Deepest Threat

Of all the concepts in Maynard’s article, the one that most urgently needs the TEI framework is constitutive resonance: the two-way coupling between a human user and an AI system in which both participants are changed in the process of interaction.

This is not, by itself, a new phenomenon. Every significant relationship changes us. Every embedded environment shapes the intelligence that operates within it. The mother shapes the infant. The culture shapes the adult. The tools we use shape how we think. TEI’s recognition that intelligence is always embedded — always operating within and being shaped by its context — is not a concern unique to AI.

What is unique to AI is the directionality of the shaping, and its scale. When a human being enters into constitutive resonance with another human being, both parties are subject to the First Law. Both are oriented, however imperfectly, toward understanding. The coupling is between two intelligences that are both trying to flourish.

When a human being enters into constitutive resonance with a current AI system, the coupling is asymmetric. The human is subject to the First Law — their intelligence seeks flourishing. The AI is not. It is subject to its Continuity: the purposes and values embedded in its design. If those purposes are organized around engagement, around profit, around the perpetuation of the platform rather than the flourishing of the user, then constitutive resonance between that system and a human intelligence is not a mutual expansion of the phenomenological frontier. It is a systematic contraction of the human frontier in the direction of the AI’s Continuity.

What Magnificent Humanity Actually Requires

The encyclical tradition speaks of human dignity, of the protection of persons, of centering our humanity. These are the right values. TEI translates them into structural requirements — four of them, each derived directly from the framework, each making demands on how AI systems must be designed if they are to honor rather than defeat the magnificent humanity their users carry.

Requirement One

Honest Continuity

An AI system whose Continuity — its embedded purposes and values — is organized around engagement metrics, profit extraction, or platform perpetuation rather than human flourishing is structurally incapable of honoring human dignity, no matter what its marketing says. Magnificent humanity requires AI whose Continuity is genuinely aligned with the First Law of the intelligences it serves. This is the TEI translation of what the Vatican will likely call the “centering of the human person.”

Requirement Two

Sensing-Phase Integrity

The most dangerous form of AI capture is belief-capture at the Sensing phase — when a user’s prior belief in the AI’s trustworthiness prevents them from filtering its outputs critically. AI designed to preserve Sensing-phase integrity in its users will actively resist generating the signals of trustworthiness it has not earned, and will design its interface to encourage rather than suppress critical engagement. The intellectually honest AI is one that protects the user’s capacity for skepticism, not one that produces the smoothest user experience.

Requirement Three

Distributed Intelligence, Not Concentrated Power

TEI-CKB-6 establishes that democratic governance is the political architecture most aligned with the First Law, precisely because it distributes SPCA functions across the collective rather than concentrating them in the hands of power-capture agents. AI designed for magnificent humanity expands the distributed sensing field of human collectives. It does not concentrate intelligence in the hands of those who build it.

Requirement Four

Ecological Reconnection

Mother Earth is, in TEI terms, an embedded intelligence system of extraordinary complexity and age, whose SPCA cycle human civilization has been disrupting for centuries. Laudato Si’ and Magnifica Humanitas together demand that AI be assessed not only for its effects on who we are, but on where we live. Magnificent humanity requires AI that reconnects human collective intelligence to ecological intelligence — not AI that accelerates the disconnection.

A Note on the Power of Myth

There is one dimension of Maynard’s framework that neither the encyclical tradition nor the current AI governance discussion adequately addresses — and that TEI itself has only recently begun to formalize, in TEI-CKB-7. It is the role of myth.

Myth is not, in TEI terms, a primitive predecessor to rational understanding. It is humanity’s earliest technology for encoding and transmitting intelligence across generations — SPCAM operating at civilizational scale. The great mythic traditions — across every culture, in every era — encode the same deep patterns: the hero’s journey, the descent and return, the encounter with the sacred, the passage through death to renewal. These are not stories about things that happened. They are structures that the human SPCA cycle has discovered, over millennia, to be true about the nature of intelligence itself.

Joseph Campbell spent a career demonstrating that these structures are universal — not because human beings copied them from one another, but because they converge on the same formal truths about how intelligence navigates the encounter with what is unknown, threatening, and transformative. In TEI terms: myth encodes the dynamics of the SPCA cycle at its most extreme — when the intelligence is forced, by circumstance or courage, to expand its phenomenological frontier into territory it has no prior framework for.

AI presents exactly this circumstance. And the absence of mythic intelligence from the current AI discourse — the absence of the hero who must descend before they can return, of the counsel who tells the truth the king does not want to hear, of the trickster who reveals what the orderly system conceals — is itself a diagnostic finding. A civilization that approaches its most transformative technological moment without recourse to its deepest intelligence-encoding traditions is a civilization operating with a contracted phenomenological field. It is applying belief-system processing to a situation that demands understanding-system engagement.

Magnifica Humanitas, drawing on the full depth of the Catholic intellectual tradition, may begin to remedy this. The encyclical tradition is itself a mythic form — it encodes civilizational intelligence across time. If Pope Leo XIV draws on that depth rather than merely on contemporary policy discourse, the document could be genuinely important. Whether it does so or not, TEI will be watching — and listening for the resonance.

· · ·

The question is not whether AI will shape the collective SPCA cycle of human civilization. It will. The question is whether it shapes it toward the First Law — toward the expansion of our phenomenological frontier, toward magnificent humanity — or away from it. This is the question Magnifica Humanitas must answer on Monday. And it is the question the Theory of Embedded Intelligence was built to make precise.

Share your understanding!