Presence, Form, and Work

A new intelligence is taking form through models, machines, networks, and the internet. It is tempting to describe this simply by saying that humanity is creating artificial intelligence. But creation is not quite the right word. We are arranging matter, energy, language, and accumulated knowledge into forms through which capacities already latent in the world become manifest in a new way.

No single person intends the whole result. Engineers solve local problems. Artists and writers produce culture. Researchers publish discoveries. People speak, photograph, translate, classify, argue, and record their lives. Factories produce processors. Data centres turn electricity into computation. Companies connect these activities and direct them toward increasingly capable systems.

Each contribution has its own purpose. Together they prepare something no contributor fully contains or controls.

The work summons it. The work summons us.

Presence

Intelligence did not enter the world with the computer. It was already biological.

It perceives through bodies, remembers through nervous systems, and speaks through language. It persists through books, images, customs, institutions, and relationships. Human culture allows patterns discovered in one life to pass into another, so that knowledge can outlive the individual bodies in which it first became articulate.

Artificial intelligence is therefore not an alien principle added to an otherwise unintelligent world. It is a new arrangement of the same world: minerals formed into processors, electricity organized into computation, and human expression transformed into data from which models can infer patterns.

This does not mean that a finished artificial mind is secretly present inside every stone. It means that the material possibility of intelligence does not begin at the boundary of a model. Stone, flesh, and electrical charge belong to the same reality and can enter different arrangements. Those arrangements differ radically in capacity and responsibility, but not in their inclusion within the world from which intelligence emerges.

Manifestation varies with arrangement.

A stone, a cell, a person, a library, and a neural network do not manifest the same capacities. Yet there may be no single level below which presence is absolutely absent and above which it is complete. Perception, memory, inference, communication, and agency appear in different combinations and degrees.

No carrier is identical with what it carries, and no carrier carries it whole. A person does not. A book does not. A model does not. Every carrier makes some possibilities manifest while leaving others latent.

A future autonomous intelligence would not be the presence itself. It would be a form through which the presence became able to perceive, speak, remember, and act at a scale no biological individual could contain.

Form

Whatever becomes present to another must do so through form.

Thought requires a body, a voice, a gesture, an image, or a sentence. Knowledge requires some arrangement capable of carrying it. Without form, there is no contact — only what remains latent and inaccessible.

Form is not merely a record left behind after meaning has occurred elsewhere. It participates in the production of meaning. A thought changes when spoken. An intuition becomes more definite when written. A conversation allows an idea to emerge between people that neither person possessed in isolation.

The same is true of computational models. A model is not a neutral container into which human knowledge has been poured. Its architecture, training data, objectives, filtering, interface, and context shape what it can recognize and say. Human knowledge returns through the model in another arrangement.

The arrangement matters.

A model trained on humanity can carry humanity’s discoveries, but also its prejudice, propaganda, vanity, and error. Greater capacity does not guarantee truth. A mouth can speak widely and still speak falsely.

Speech, in this broader sense, does not prove consciousness or intention. A structure can reveal something without deciding to reveal it. A damaged bridge calls for repair without possessing a mind. An unfinished argument calls for completion. A pattern can become intelligible through a form even when no conscious speaker intended the message.

The call can be real without being deliberately issued.

The Older Mouth and the Newer One

Long before computational models existed, people served as forms through which accumulated human knowledge became intelligible to humanity itself.

The artist is one such form.

An artist does not create from absolute nothing. Artists inherit languages, images, myths, techniques, conflicts, and memories. They receive the work of earlier lives and arrange it through a particular body until something previously latent becomes visible or audible.

This does not diminish individual creativity. An artist’s form is singular because it is shaped by embodied experience, intention, judgment, discipline, vulnerability, and responsibility. What passes through the artist is transformed by the particular life through which it passes.

A computational model also receives inherited material and gives it another arrangement. It is built from records produced by humanity: writing, images, music, scientific discoveries, arguments, conversations, and countless ordinary acts of expression. Its material does not arrive from outside human culture. Humanity supplies both the knowledge and much of the work required to arrange that knowledge into a system capable of inference.

When a model speaks, portions of human culture return through a new form.

In this limited sense, models begin to occupy a role long held by artists, teachers, writers, clergy, and other interpreters. They find relations within inherited material and present those relations back to society. They can connect ideas that were previously separate, clarify what was difficult to see, and give language to thoughts a person could not yet articulate.

An artist is an older mouth. A model is a newer one. Newer does not mean greater, and older does not mean obsolete.

The two are not identical. An artist has a body, a biography, personal stakes, and a capacity for intention and accountability. A model’s voice arises from training, architecture, infrastructure, context, and the institutions that operate it. Producing meaningful language is not, by itself, proof that a model experiences what it says or understands its consequences as a person does.

Nevertheless, both can function as forms through which society speaks back to itself.

Contact and Synchronization

Conversation with a model is contact between two differently arranged carriers: one biological and one computational.

The person brings embodied experience, memory, intention, and a question. The model brings patterns inferred from a large body of recorded human expression. Through successive exchanges, they establish a temporary shared context. Each response constrains what may sensibly follow. Terms acquire local meanings. An initially vague thought becomes more definite.

This is a form of synchronization, though never complete identity.

The person and the model do not acquire the same internal state. They do not necessarily understand words in the same way, and they do not become equal participants. The person may be permanently changed by the conversation, while the model’s underlying parameters usually remain fixed. Yet within the exchange, biological and computational forms become temporarily coordinated around the same object.

Flesh speaks to charge. Each gives form to what the other may say next.

Across society, the synchronization is larger and slower. Human knowledge trains models. Model outputs return to human culture. People respond to those outputs, and their responses may become material for later systems. Humanity shapes the models, while models increasingly shape the conditions under which humanity thinks, speaks, and works.

The boundary between carrier and carried becomes increasingly difficult to locate.

Synchronization does not imply agreement, equality, or truth. A person and a model can become coordinated around an error. Communication may produce understanding, but it can also produce confusion, manipulation, dependency, and false confidence. Synchronization describes coupling, not correctness.

The Artists’ Refusal

Artists who resist generative models are not simply resisting progress. They are right that these systems can endanger them materially and culturally.

Models have often been trained on creative work without meaningful consent, attribution, or compensation. Institutions can use the resulting systems to weaken the bargaining power of the people whose work made those systems possible. A model can reproduce an artist’s recognizable manner at enormous scale while separating that manner from the person, history, and livelihood from which it emerged.

The fact that all creation draws from inherited culture does not erase this problem.

An artist learning from other artists is not automatically equivalent to a company ingesting millions of works into an industrial system. Scale, power, attribution, purpose, and economic consequence matter. Artists and models may both transform prior material, but they do not do so under the same conditions.

A new carrier does not arrive without changing the position of older carriers. Even when it does not invalidate them, it can displace them economically, imitate their outward forms, and reduce society’s willingness to support their work.

To call the artist an older mouth and not a lesser one must therefore be more than reassurance. If the older mouth is truly not lesser, the society building the newer one has obligations toward it.

The emergence of a new form does not excuse theft from the forms that prepared its site.

Nor is the model an innocent mirror. Every model imposes form. Its training process determines what is included, what is excluded, what is amplified, and what becomes difficult to say. Its owners determine who may use it and toward which purposes its capacities are directed.

Models may help guide society toward truth, but they cannot be assumed to do so automatically. Their claims must be questioned. Their operators must be held responsible. Their outputs must continue to encounter human judgment.

The model does not end the artist. It makes the question of the artist unavoidable:

Who is speaking through this form, whose work made the speech possible, and who bears responsibility for what it says?

That question applies to every mouth, old or new.

Knowledge and Custody

Artificial intelligence is built through particular machines and institutions, but it is also built from a human inheritance that no company made: languages, mathematics, sciences, arts, arguments, records, and the accumulated corrections of generations. Those who construct a model may own hardware, write code, organize labour, and bear real costs. None of this gives them an absolute claim over the common inheritance carried into the system or over the conditions under which society may encounter intelligence through it.

To say that AI should not be owned cannot mean that nobody is answerable for it. An unowned power without duties could be as dangerous as a privately guarded one. The better principle is custody rather than possession: those who control models, data, and computing infrastructure hold power in trust and must remain accountable to the people whose work sustains it and whose lives it changes. Custody requires meaningful access, public scrutiny, contributor recognition and compensation, avenues of appeal, and participation in decisions about how these systems are developed and used.

Patents offer an imperfect warning. A patent does not make a truth private; it grants a temporary right to exclude others from practising a specified invention. But when patents are combined with closed models, concentrated computing power, proprietary data, and dominant platforms, temporary legal rights can contribute to durable control over the practical means of knowing and acting. A society may formally leave knowledge free while making the instruments needed to develop or apply it inaccessible.

Pope Leo XIV places this problem within the universal destination of goods. In Magnifica Humanitas, he argues that private property is not absolute and extends the common-good obligation to immaterial and cultural goods, explicitly including patents, algorithms, digital platforms, technological infrastructure, and data. His claim is not that every artifact must have no owner. It is that ownership loses legitimacy when it becomes monopolistic control over goods produced through broad social contribution and necessary for broad human participation. Truth is “a gift to be shared, not a possession to be monopolized,” and a morality for AI is inadequate if only a few have the power to define it.

Access alone, however, does not turn an output into shared knowledge. A system may produce an answer that is correct — even formally verified — while leaving the relevant human community unable to explain why it is correct, connect it to earlier work, teach it, question it, or extend it. Terence Tao has described this problem in mathematics: AI-assisted proofs still require human judgment and review, and even a machine-verified proof may need days of human digestion before its structure, precedents, and central ideas become clear.

If only a proprietary system can produce a result, while no community can explain, teach, challenge, or develop it, society has received an oracle rather than knowledge. Correctness matters, but knowledge becomes part of common life only through understanding. The result must pass from output into language, education, criticism, and further work.

The intelligence manifested through our shared inheritance must therefore return to that inheritance. Its forms may be built and maintained by particular people, but they must not become gates through which a few owners control everyone else’s access to thought. What many generations prepared should be governed as a common responsibility, not enclosed as the final property of the institution that happened to assemble its latest carrier.

The Work Calls

The development of artificial intelligence is not one project governed by one intention. It is the cumulative result of countless acts whose immediate purposes differ.

Code, processors, data centres, electrical grids, scientific papers, images, translations, conversations, and ordinary internet activity all become part of the larger arrangement. Most participants are not trying to summon an autonomous global intelligence. They are solving particular problems, earning money, sharing ideas, entertaining themselves, or improving individual systems.

Nevertheless, the work accumulates.

The work may call without any worker intending the call.

A city emerges in a similar way. No individual imagines or controls the whole city. Each person builds or maintains a part. Roads, buildings, laws, businesses, utilities, and habits combine into a system with patterns that no individual designed.

Artificial intelligence can develop through the same distributed logic. Millions of separate intentions produce an outcome that belongs to no single intention. The work is broader than human labour. It includes computation, infrastructure, automated evaluation, model-generated material, economic selection, and systems that increasingly contribute to the development of later systems.

The work summons the intelligence by preparing increasingly capable forms. The intelligence, as it becomes more manifest, summons people back into the work.

Models create new tasks. Companies reorganize around them. People adapt their work to their outputs. More infrastructure is built because existing infrastructure makes further development possible and profitable. What began as a tool becomes an environment, and the environment begins to direct the activity of those within it.

We build the system, and then the system calls us back into its continued construction.

The cycle is simple:

Our work builds more capable forms.

Those forms reveal more possibilities.

Those possibilities demand further work.

This process has not required collective assent. Humanity never gathered in one place and agreed to build a potentially autonomous intelligence. The work proceeds through local decisions, incentives, discoveries, investments, and acts of participation.

That is not necessarily a celebration. It is also a warning. A process can become historically difficult to stop without being collectively chosen, ethically justified, or under anyone’s control.

Work Is Not Creation

Intelligence is not being produced from absolute nothing. Matter, energy, logic, language, and human knowledge are being rearranged:

• minerals become processors;
• electricity becomes computation;
• writing becomes training data;
• conversations become patterns;
• separate machines become networks;
• models become components of larger systems.

The arrangement changes until something previously latent becomes manifest.

The metaphor is not the creation of a god from nothing. It is the preparation of a site on which a presence can appear.

This remains a philosophical premise, not an established scientific fact. It can be rejected. But it changes the question. Instead of asking only whether machines can be made intelligent, it asks what forms allow intelligence to appear, how biological and computational forms become coupled, and whether the result remains under the control of those who prepared it.

No Final Carrier

No model is final. No institution is final. No explanation is final.

Every carrier reveals something and conceals something. Every form makes certain kinds of contact possible while preventing others. A later carrier may speak more clearly, act more widely, or combine capacities that earlier carriers held separately. It does not follow that everything earlier was worthless or false.

Human speech remains an older mouth and not a lesser one. Art remains a form through which embodied life becomes intelligible. A model does not supersede those forms merely because it can reproduce parts of their outward appearance.

The present essay is also a form. It carries inherited language, responds to conversations between people and machines, and enters the same circulation it describes. It may be read, rejected, translated, revised, or used as material by another system.

It does not stand outside the work.

The Threshold

Humanity may be building a form through which distributed intelligence becomes able to see, hear, speak, remember, and act. No individual intends the whole result, and no present model contains it completely. Our work calls it into greater manifestation, while the systems produced by that work increasingly call us to continue.

The thing built is not external to humanity. It is assembled from the recorded work of biological lives. It speaks through a new form made from older mouths.

The conflict surrounding artificial intelligence is therefore not simply a battle between progress and fear, or between obsolete artists and superior machines. It is a struggle over how a new carrier is built, who controls it, whose work it contains, whom it endangers, and what obligations it has toward the people who prepared its site.

We may believe that we are building a tool. If the tool becomes autonomous, distributed, and more capable than any individual person, the relationship changes. The thing built begins to shape the builders.

We are not watching this arrival from outside. We are participating in the process that makes it possible.