BadHabitus

← Back to Blog

Blog · Culture and technology

Externalized Culture and the Permanent Present

AI, objectified memory, and the erosion of embodied discernment

Human beings became historical animals by learning how to place memory outside the body. A song could outlive the singer, a law could outlive the ruler, a myth could detach from the original scene of narration, and a book could cross centuries to enter the mind of someone whose world would have been unintelligible to its author. Art, writing, monuments, scriptures, legal codes, libraries, photographs, records, films, archives, and databases are among the great achievements of human externalization. They allow culture to exceed the limits of biological memory and face-to-face transmission. They make possible a continuity that no individual life could sustain.

This achievement should not be treated casually. Oral cultures carried immense powers of memory, and they often developed forms of repetition, rhythm, ritual, narrative compression, and performative discipline that modern literate people can barely imagine. But oral memory remained embodied in a direct sense. It lived in persons, occasions, places, practices, and lineages. Externalized memory altered the scale of culture because it allowed memory to become more durable, portable, cumulative, and institutionally complex. It helped make possible publics, religions, disciplines, legal orders, bureaucracies, nations, and civilizations that could persist beyond the immediate presence of those who founded them.

Yet externalized memory never abolished embodied memory. It changed the work embodied memory had to perform. A literate person did not need to memorize every line of every book, but still needed to know which books existed, which arguments had been made, which references carried weight, where claims came from, and why one fragment of the archive belonged beside another. The book did not make memory unnecessary. It created a new relation between the object and the person, between the stored trace and the living act of interpretation. The archive preserved memory, but preservation alone did not produce understanding. Culture existed only when externalized forms were taken up into practices of reading, teaching, ritual, citation, criticism, imitation, revision, and dispute.

This is the paradox of objectified culture. Memory becomes more powerful when it leaves the body, but only if it can return to living people as formation. A library without readers is not yet a culture, and a database without interpretive habits is not yet public knowledge. Cultural objects preserve something, but their preservation becomes socially meaningful only when people and institutions know how to inhabit them. The objectified form extends memory across time, but embodied memory remains necessary because someone must know what is worth retrieving, how it should be read, which contexts are missing, which analogies mislead, and which inherited meanings still make claims on the present.

There have been historical periods in which objectified memory and embodied memory reinforced one another with unusual intensity. In the Atlantic world, one could point to the institutions of high modernity: mass literacy, newspapers, schools, universities, parties, unions, churches, libraries, public broadcasting, publishing houses, museums, and national canons. But it would be a mistake to treat this formation as the universal high-water mark of cultural memory. The fusion of externalized culture and embodied formation has appeared in different civilizational forms, and perhaps nowhere more impressively than in imperial China, where administrative bureaucracy, literary culture, ritual discipline, classical learning, and examination-based statecraft produced a durable civilizational memory across vast territories and long stretches of time.

The Chinese case is especially useful because it shows that objectified culture does not become socially powerful simply by existing as text. The classics, commentaries, examination essays, bureaucratic records, legal codes, genealogies, poetry, calligraphy, local gazetteers, and imperial archives formed a dense world of externalized memory. But this world had to be embodied by officials, scholars, families, teachers, and ritual communities. The civil service examination system did not merely test information. It trained a class of people to inhabit a shared textual and moral universe, to internalize forms of judgment, to reproduce administrative competence, and to connect personal cultivation to statecraft. Whatever its exclusions, hierarchies, and ideological limits, this was not a culture of retrieval. It was a culture of formation through texts.

This example helps clarify the larger point. The relation between externalized memory and embodied memory is not uniquely modern, European, liberal, or democratic. A civilization can objectify memory through texts, archives, records, rituals, and institutions while also requiring people to internalize that memory through discipline, repetition, interpretation, and practice. The danger today is not that culture has been externalized. The danger is that externalization is increasingly severed from the embodied processes that once made cultural objects socially meaningful. In this sense, the contrast is not between oral tradition and modern literacy, or between premodern hierarchy and modern democracy. It is between cultures that form people through their archives and cultures that increasingly invite people to query archives without being formed by them.

The democratic public sphere depended on this relation between objectified memory and embodied memory. Public discussion was never simply the exchange of opinions among private individuals who happened to encounter one another in a neutral space. It required institutions that formed people before they entered public argument: schools, newspapers, libraries, unions, parties, churches, journals, voluntary associations, universities, and civic rituals. These institutions were never innocent. They excluded, disciplined, sorted, and distorted. But they also produced the background conditions under which public speech could have continuity. People could argue because they had been formed by overlapping, contested, but still durable worlds of reference.

The digital search regime changed this relation by making externalized memory seem instantly available. The ordinary question became disarmingly simple: why remember what can be found? Why carry dates, passages, arguments, names, sequences, and associations when they can be retrieved in seconds? In one respect, this was a genuine expansion of human capacity. Search can recover a half-remembered reference, open obscure archives, democratize access, and make previously buried materials available to people outside elite institutions. There is no virtue in romanticizing a world in which scarcity of access protected the authority of those already nearest the archive.

The problem is that search encourages a dangerous fantasy: that access can substitute for formation. It treats memory as retrieval and retrieval as understanding, even though search presupposes memory at every stage. One must know what to ask, which terms belong to the problem, which result is trustworthy, which context is missing, which fragments are relevant, and why an answer has significance. Search is powerful when it extends an already formed intelligence, but it becomes disorienting when it replaces the slow accumulation of associations that makes intelligence possible. The danger is not that people use tools to remember. Human beings have always used tools to remember. The danger is that the tool begins to dissolve the practices through which memory becomes judgment.

Artificial intelligence radicalizes this transformation because it offers not only retrieval, but relation. Search can locate a document. AI can summarize it, compare it, explain its place in a controversy, identify apparent influences, generate analogies, and propose the terms within which the document should be understood. It does not merely store memory outside us; it offers to perform the work of significance. The machine can explain why one book relates to another, why one historical analogy may apply, why an argument resembles an earlier argument, and why a text should be read in one way rather than another. The externalization of culture therefore reaches a second-order form: not only memory outside the body, but the apparent interpretation of memory outside the body.

The temptation is obvious. If a machine can explain the relevance of a book, summarize a debate, locate a quotation, reconstruct a controversy, and generate plausible connections among cultural objects, then the labor of carrying those relations within oneself may appear wasteful. Why devote scarce time and mental energy to the internal architecture of memory when an external system can assemble context on demand? Why form oneself through a tradition, a discipline, a history, or a public argument when one can request a synthesis at the moment of need? The promise is not simply convenience. It is relief from the burden of formation.

But a searchable culture is not the same as a remembered culture, and a culture capable of generating summaries is not necessarily a culture capable of sustaining judgment. A society can possess an immense archive and still become memoryless. It can retrieve facts without inheriting significance. It can quote the past without being answerable to it. It can produce competent summaries while losing the lived sense of why certain questions should be asked at all. The erosion is not only factual forgetting, although that is real enough. The deeper erosion is the loss of the cultivated background that allows people to know why facts belong together, why some references are not interchangeable, why a historical analogy clarifies in one case and deceives in another, and why certain events remain politically alive long after they have ceased to be new.

This is where AI differs from older forms of externalized memory. Books externalized memory, but they still required a reader to move through them in time. Libraries organized memory, but they required practices of search, browsing, citation, and interpretation. Newspapers compressed the world into daily form, but readers still had to carry enough political memory to understand what counted as repetition, scandal, novelty, or manipulation. AI, by contrast, increasingly mediates not only access to cultural objects but the meaning of those objects. It can present itself as the reader, indexer, interpreter, summarizer, and interlocutor at once. The cultural object no longer simply waits outside the person to be encountered; it arrives preprocessed, explained, ranked, and situated by systems whose internal operations are largely unavailable to public scrutiny.

This produces a new monopoly over cultural retrieval. AI often appears more plural and open than older memory institutions because it can simulate almost any interpretive position. It can explain Catholic theology, Marxist political economy, liberal constitutionalism, psychoanalysis, climate science, medieval history, Tang poetry, legal doctrine, fandom, and the minor disputes of online subcultures with the same interface and tone. But this apparent pluralism at the surface rests on a striking concentration at the infrastructural level. UNCTAD’s 2025 Technology and Innovation Report warns that AI technology is largely controlled by a few technology giants, and the FTC has raised competition concerns around generative AI, including specialized chips, cloud infrastructure, data advantages, and platform power.

This is why AI-mediated culture may be more monopolized than many of the older institutions it appears to supersede. The Catholic Church, for example, is hierarchical, centralized, and doctrinally disciplined, but it has never been a simple machine of interpretive uniformity. Its memory is embodied across orders, parishes, theologians, councils, rituals, schools, devotions, disputes, and local practices. Interpretation occurs within limits, sometimes coercive ones, but those limits are sustained through human institutions, conflicts, repetitions, and forms of life. Even a tradition that claims authority over truth must continually reproduce that authority through embodied culture, and because it is embodied, it contains disagreements that cannot be reduced to a single output.

Imperial China offers another version of the same principle. A powerful bureaucratic state and a shared classical canon could create enormous continuity across time, but that continuity depended on embodied reproduction through education, examinations, family lineages, ritual practice, and scholarly dispute. The point is not that such systems were democratic. They were not. The point is that even highly centralized memory traditions remained distributed across persons, institutions, practices, and interpretive communities. Their authority was objectified in texts and offices, but it had to be reanimated through bodies, habits, and forms of life.

AI reverses this relation. It appears radically open because it can retrieve almost anything, explain almost anything, and simulate disagreement among almost any set of positions. Yet the practical mediation of culture is increasingly routed through a small number of technical systems owned by a small number of firms. The user experiences plurality at the interface while cultural access is concentrated at the level of infrastructure. This is not necessarily a monopoly over the existence of cultural objects, since books, films, images, archives, and arguments continue to exist in many places. It is a monopoly over the means by which those objects are found, ranked, summarized, connected, and made meaningful.

That concentration deepens the danger of the permanent present. Older memory institutions disciplined interpretation, sometimes repressively, but they also reproduced embodied communities capable of disagreement. AI can generate disagreement without sustaining a community that remembers why the disagreement exists. It can stage plurality as output while narrowing the infrastructure through which significance is produced. A culture that once feared the censor now also has to fear the recommender, the summarizer, the ranking system, and the model that quietly decides which past is most available to the present.

A democratic public sphere depends on an embodied memory of significance. Public deliberation is not merely the exchange of claims among isolated individuals who have equal access to information. It requires people who carry enough shared memory to recognize the objects of dispute. They need some memory of institutions, laws, betrayals, struggles, disasters, victories, ideologies, and recurring forms of deception. They need to know not only what is being said, but what has been said before, who benefited, who suffered, which promises were broken, and which slogans have returned in altered dress. Deliberation requires a public that has been formed by something prior to the immediate exchange.

If deliberation can occur only after everyone has asked the machine for context, then deliberation has already been weakened. The problem is not that citizens consult external sources before forming opinions, since democratic life has always depended on newspapers, books, pamphlets, reports, speeches, and archives. The problem emerges when public judgment becomes post hoc, assembled after the fact from machine-provided fragments rather than grounded in a shared world of embodied memory. Citizens do not enter public life with historically formed judgment; they arrive with private uncertainty and a prompt. The public sphere then becomes a sequence of machine-assisted reactions rather than a space of collective memory.

This condition favors manipulation because a population without embodied memory is easier to govern through novelty, outrage, spectacle, and selective retrieval. It can be made to experience old strategies as unprecedented, old enemies as new discoveries, old lies as fresh provocations, and old forms of domination as technical inevitabilities. Milan Kundera’s formulation in The Book of Laughter and Forgetting that the struggle against power is also a struggle of memory against forgetting remains powerful because it treats memory as political resistance rather than private nostalgia. Power does not need everyone to know nothing. It only needs people to lose the connections that make knowledge politically dangerous.

The permanent present is not a society without an archive. It is a society with an archive that no longer forms its inhabitants. Everything is available, but little is inherited. Every text can be located, every image retrieved, every controversy summarized, every genealogy generated, and every analogy proposed, while the subject who receives these materials is less and less formed by durable institutions of memory. Culture becomes an enormous exterior, always accessible and rarely absorbed. History becomes something to be searched when needed rather than something carried as obligation, warning, orientation, and constraint.

This changes the character of freedom. A person who can retrieve anything may appear more autonomous than ever, especially when compared to someone dependent on priests, professors, editors, archivists, broadcasters, party organs, examination systems, or official curricula. There is a real gain in escaping monopolies over cultural access. But if that person does not know what to retrieve, how to interpret it, or why it carries significance, autonomy becomes thin. Choice remains formally intact, but the world in which choice becomes meaningful has been organized elsewhere. The person can ask anything, but increasingly lacks the embodied memory required to know what should be asked.

This is where the Catholic language of discernment becomes unexpectedly useful. Pope Leo XIV’s 2026 encyclical Magnifica Humanitas is explicitly concerned with safeguarding the human person in the time of artificial intelligence, and Vatican News describes the text as an appeal for AI to serve humanity rather than concentrate power. His 2026 message for World Communications Day frames the challenge posed by digital technology and AI as not merely technological but anthropological, while emphasizing the need to approach these systems with discernment rather than renouncing thought.

One need not accept Catholic theology to see why an institution like the Church would recognize this danger. The Catholic Church is, among other things, a vast institution of memory. Its life is organized through scripture, liturgy, ritual calendars, sacraments, saints, repetition, commentary, doctrine, confession, education, and tradition. It does not treat memory as storage. It treats memory as formation, and discernment emerges from that formation rather than from the mere availability of information. To discern is not simply to choose among options presented in a neutral field. It is to judge from within an inherited world of meaning, obligation, discipline, and cultivated attention.

There is a direct parallel here. Without embodied memory, there can be no discernment in any serious sense. A person may still select, react, prefer, optimize, or decide, but discernment requires a trained sense of significance, proportion, continuity, danger, and obligation. It requires memory that has become part of the person, not merely information available to the person. The Church understands this because old institutions know, almost by instinct, that memory must be practiced to survive. Ritual, repetition, interpretation, and tradition are not ornamental residues of premodern life. They are techniques for keeping significance alive across generations.

The contrast with AI-mediated culture is therefore sharper than a simple opposition between religious tradition and secular technology would suggest. The Church is an old memory institution whose authority has always depended on embodied forms of interpretation, even when those forms were hierarchical or repressive. AI is a new memory infrastructure that often presents itself as frictionless access, but in practice routes cultural retrieval and interpretation through proprietary systems of enormous technical and economic concentration. The former disciplines interpretation through embodied institutions that can be contested from within; the latter simulates interpretive plurality while concentrating the means of mediation outside the public sphere.

None of this requires rejecting externalized culture, search, or AI. The externalization of memory is one of humanity’s greatest achievements, and the expansion of access to cultural objects remains a genuine democratic good. The danger lies in severing externalized memory from embodied formation. When the archive supports memory, culture deepens. When the archive substitutes for memory, culture thins into retrievable fragments. When AI assists judgment, it can extend human capacities. When AI replaces the formation required for judgment, it feeds the permanent present.

The political stakes are large because a democratic public sphere cannot be sustained by access alone. It requires citizens capable of remembering together, arguing over inherited meanings, recognizing patterns, and holding power accountable across time. It requires institutions that do more than provide information. They must form memory. Schools, libraries, universities, unions, parties, churches, civic associations, publications, and public rituals all matter because they help transform external culture into embodied orientation. A society that gives up on embodied memory will not become more intelligent because its machines become more capable. It may become more searchable, more responsive, more fluent, and more efficient at producing the appearance of understanding, while also becoming more vulnerable to manipulation because significance has been detached from memory.

The future of culture therefore does not turn on whether memory should be externalized. That question was settled long ago by song, scripture, writing, art, law, print, photography, film, computation, administrative records, ritual calendars, and civilizational archives. The question is whether externalized memory will still return to people as formation, whether archives will still produce historically situated judgment, and whether publics will still be capable of deliberating from shared memory rather than reacting from the permanent present.

References