When everyone can generate their own music, movies, games, books and worlds with AI, the rarest form of media may become something we experience together.

For most of modern history, culture has been shaped as much by scarcity as by creativity. There were only so many television networks, movie screens, radio stations, record labels, publishing houses and shelves in the local bookstore. Only a fraction of the music recorded in any given year received national distribution. Only a fraction of the films imagined were actually made. Even when artists wanted to create something, the economics of production and distribution imposed a brutal question: Is there a large enough audience to justify making this?

Those limitations created powerful gatekeepers, and there is little reason to romanticize them. Studios, publishers, networks, labels and retailers possessed enormous influence over which voices reached the public and which disappeared. Yet those same bottlenecks produced something valuable almost accidentally. Because relatively few cultural artifacts could reach enormous audiences, millions of people repeatedly encountered the same ones.

That gave us shared experience.

People watched the same television finales, waited for the same movies, heard the same songs and recognized the same celebrities, advertisements, jokes and stories. You could walk into work after a major television event and reasonably expect someone else had watched it. Put a familiar song on at a party and an entire room might recognize the opening notes. Quote a movie and someone might finish the line.

Mass media did more than distribute entertainment. It manufactured common cultural territory.

Artificial intelligence may be about to blow that territory apart.

The End of Content Scarcity

Generative AI is rapidly reducing the cost and technical difficulty involved in producing almost every form of media. Images came early. Writing followed. Music, voices and increasingly sophisticated video arrived behind them. Games and interactive environments are moving along the same trajectory. None of these technologies is finished, and today's limitations are substantial, but the direction is difficult to miss.

Eventually, the distinction between finding entertainment and creating entertainment may become surprisingly thin.

Instead of searching a streaming service for a science-fiction series, you might describe the series you want: a slow political drama aboard a generation ship 300 years from Earth, historically informed, morally ambiguous and twelve episodes long. Rather than buying the latest open-world game, you might request an alternate-history RPG in which the Roman Empire survived into the industrial age, complete with realistic economics, political factions and characters capable of remembering years of interaction.

Music could become even more personal. Instead of searching for a playlist that helps you concentrate, an AI system might understand precisely what kinds of rhythms, harmonies, tempos and textures help you work and generate an endless soundtrack that has never existed before and may never exist again. Family photographs might become interactive environments. A grandfather's journals could become a film. A favorite fictional universe could continue indefinitely because there would no longer need to be a final book, episode or game.

Humanity may be approaching something that previous generations could scarcely imagine: effectively infinite media.

And infinite media creates a peculiar problem. If almost anything can exist, what does everyone know?

Eight Billion Channels

The fragmentation of mass culture did not begin with artificial intelligence. Cable television fractured the television audience. The internet fractured it further. Streaming replaced a relatively small number of schedules with enormous on-demand libraries. Social media then personalized the information environment itself, producing feeds that increasingly differ from person to person.

Generative AI represents another step, but it may be a qualitatively different one.

Today's recommendation algorithms choose among things that already exist. Tomorrow's systems may create the thing being recommended. Instead of selecting the video most likely to hold your attention, the system could generate a video specifically designed to hold your attention. Instead of identifying the novel you are most likely to enjoy, it could write one. Instead of recommending a game, it could construct one around your preferences.

The ultimate recommendation engine does not recommend. It creates.

Consider where that leads. Your movie can have your preferred pacing, aesthetic and emotional intensity. Your game can contain the mechanics and difficulty you enjoy. Your fictional universe can emphasize the themes that fascinate you. Stories can avoid subjects you dislike or, more troublingly, ideas that challenge you. Characters can become exactly as comforting, provocative, funny or sympathetic as the system has learned you want them to be.

The age of mass media gave us a handful of channels. The internet gave us millions.

Generative AI could give us eight billion.

Cultural Loneliness

Now imagine watching one of the greatest movies you have ever seen. It moves you deeply. The performances are extraordinary, the story seems uncannily relevant to your life and the ending stays with you for days. Naturally, you want to talk about it.

But nobody else has seen it.

Nobody else ever will.

The movie was generated specifically for you, perhaps in real time, and it may not even exist anymore.

That represents a fundamentally different relationship with culture because part of the meaning of art has always existed outside the artifact itself. A song matters partly because of what happens between you and the song, but it can also matter because you remember singing it with someone twenty years ago. A film becomes part of culture because millions of strangers remember the same scene. A championship becomes mythology because an entire city remembers exactly where it was when the final seconds disappeared from the clock.

Human beings don't merely experience things. We also experience the knowledge that other human beings experienced them.

There may be a peculiar loneliness waiting inside personalized media abundance. It would not be loneliness caused by having nothing to watch, read, hear or play. Quite the opposite. We could have an inexhaustible supply of astonishing entertainment and still discover that something important is missing because so much of it belongs only to us.

We might call this cultural loneliness: isolation not from a lack of experience, but from an abundance of experiences that cannot easily be shared.

The Strange Return of Scarcity

This creates one of the more delicious paradoxes of the AI era. For centuries, communications technology has relentlessly attacked scarcity. Printing presses made books reproducible at scale. Recording technology separated music from the physical presence of musicians. Radio and television transmitted performances across continents. The internet drove the marginal cost of copying information toward zero. Streaming placed libraries that once would have seemed imperial in scale inside a device carried in a pocket.

Generative AI attacks the remaining bottleneck: creation itself.

And after spending centuries eliminating scarcity from media, we may begin deliberately rebuilding some of it.

Not because people will suddenly develop nostalgia for expensive books or limited television schedules, but because scarcity can produce commonality. A cultural artifact experienced by everyone in the same form becomes a meeting place. The inability to customize it may become part of its value.

It is possible to imagine future media advertising something that sounds completely unremarkable today: Everyone gets the same version. The story does not adapt to your preferences. The protagonist does not become more sympathetic because the system senses you are losing interest. The ending cannot be regenerated because you disliked it. This movie is this movie. This album is this album. This book ends where its creator decided it ends.

In a world of infinite personalization, limitation itself could become meaningful.

Reality Gets a Premium

This may also make certain forms of culture considerably more valuable, and sports provide perhaps the clearest example.

Nobody knows how the game ends. Nobody can personalize the score. Your AI cannot quietly rewrite the fourth quarter because you would prefer the Packers to win. Millions of people experience the same event simultaneously, and reality stubbornly refuses to optimize itself for any of them.

That stubbornness may become extraordinarily valuable.

The same principle applies to concerts, festivals, theater, religious ceremonies, community gatherings, political events, universities, restaurants and public spaces. They are places where something happens in the presence of other people and where the outcome cannot be perfectly generated for every participant.

As synthetic experience becomes more convincing and more abundant, we may assign greater value to experiences precisely because they are unscripted, shared and difficult to reproduce. A concert matters because you were there. A football game matters because everyone saw the same impossible catch. A neighborhood festival matters because the rain actually started at four o'clock and everyone ran beneath the same tent.

Reality may acquire a premium.

Place Becomes an Anchor

Place may become particularly important because physical geography provides something personalized media cannot easily dissolve: a common reference point.

A town, neighborhood, river, university, sports team, landscape or building can belong to thousands of individual stories while remaining recognizably the same place. AI can generate infinite interpretations of Wisconsin, but it cannot manufacture the historical fact that generations of people have stood beside the same Fox River, driven the same roads, attended the same schools or watched the same landscape change.

Technology can interpret place endlessly. Place itself remains stubbornly common.

This could have an unexpected cultural consequence. Hyperlocal media, historically constrained by economics, may flourish precisely because AI destroys the requirement that culture appeal to enormous audiences before it becomes economically practical to create.

A traditional record company has little reason to finance an album about the history of a small Wisconsin town. The addressable market is absurdly small by conventional entertainment economics. A documentary studio probably cannot justify spending hundreds of thousands of dollars telling that town's story either.

AI changes the arithmetic.

A community of 15,000 people can have an album about its history. A neighborhood can have a documentary. A high school can have an interactive archive. A family can have a film. A university can have music tracing its history. A village can possess cultural artifacts sophisticated enough to once have required teams of professional creators and substantial capital.

This is one of the genuine promises of generative AI. It does not merely threaten mass culture. It can make previously uneconomic culture possible.

Instead of one national cultural commons serving 100 million people, we could develop millions of smaller ones rooted in communities, interests, histories and places.

Not mass culture.

Microculture.

From Shared Consumption to Shared Creation

There is another possibility worth considering: perhaps we are looking for shared experience in the wrong place.

For most of the industrial media era, shared culture primarily meant consuming the same artifact. We watched the same television program, listened to the same record, read the same bestseller or played the same game. The creators were relatively few and the audience was enormous.

Generative AI could radically increase the number of people capable of making sophisticated cultural artifacts. That does not necessarily mean everyone suddenly becomes an artist, any more than smartphone cameras turned everyone into a professional photographer. But it dramatically lowers the technical threshold between imagining something and producing a version of it.

That could change the cultural conversation from “Did you see that?” toward “What did you make?”

People may exchange generated songs. Families may create games together. Children may construct fictional universes with friends. Communities may collectively produce histories. Fans may inhabit collaborative worlds that evolve through the contributions of thousands of participants.

The artifact remains important, but the act of creation itself becomes social.

The twentieth century was dominated by shared consumption. The AI era may introduce much more shared creation.

That could produce remarkable new forms of human expression. It could also produce another form of fragmentation if everyone retreats into private creative universes. Both outcomes can be true simultaneously.

The Problem of the Perfect Mirror

The danger of personalized media extends beyond loneliness. Shared culture has always forced us, however imperfectly, to encounter other minds.

An author does not know exactly what you want. A filmmaker makes choices you dislike. A musician releases an album that goes somewhere you did not expect. A game frustrates you. A story presents a character whose worldview irritates you. Art occasionally bores, annoys, challenges or confuses us.

That friction matters because another person made choices that were not optimized around our preferences.

A perfectly personalized AI entertainment system could gradually eliminate that friction. The machine learns what makes us laugh, what makes us angry, which characters we identify with, which subjects hold our attention, which ideas make us uncomfortable and which endings leave us satisfied. Every interaction provides another piece of information. The system improves.

Eventually, personalized culture could function less like a window through which we encounter another mind and more like an extraordinarily sophisticated mirror reflecting our own preferences back at us.

That is not merely an entertainment problem. It is potentially a social one.

Communities require common reference points. Relationships require experiences that exist independently of either participant. Democratic societies require at least some shared conception of reality before their members can meaningfully disagree about what should be done about it.

A civilization composed of billions of personalized cultural realities may discover that infinite choice carries an unexpected price.

Humans Will Build Campfires Again

None of this means shared culture disappears. Human beings are too social for that. The more likely outcome is that we invent new mechanisms for gathering attention and rediscover old ones that suddenly become more valuable.

Some cultural campfires will remain enormous: the World Cup, the Olympics, major elections, championship games, global concerts, scientific breakthroughs and historic events. Others will be tiny: a neighborhood festival, a university tradition, a group of friends sharing an AI world, a local album, a family film or a community gathering around the history of the place where they live.

The technological campfire changes. The human impulse to gather around one probably does not.

What changes is that gathering may increasingly become a choice.

For much of the industrial media era, shared culture was partly an accident of technological limitation. There simply weren't enough channels for everyone to disappear into an individually constructed universe. The limitations of production and distribution forced us into common cultural spaces whether we consciously valued them or not.

AI removes many of those limitations.

Future generations may therefore have to deliberately create common spaces, preserve common stories, attend common events and occasionally seek experiences that have not been optimized specifically for them. Shared culture may stop being something the media system automatically produces and become something communities actively maintain.

That may be one of the more important cultural adaptations required by artificial intelligence.

The Rarest Thing in the Infinite Library

There is an extraordinary upside to all of this that should not disappear beneath the warnings.

AI could allow millions of stories to exist that previously would never have been told. Communities too small to interest a record label can have music. Children can build games. Families can create films. Forgotten places can have interactive histories. People without the money, equipment, connections or technical skills once required to participate in sophisticated media production can move much closer to the creative process.

That democratization matters. Something genuinely new is becoming possible.

But technological abundance has a habit of creating new forms of scarcity. When information became abundant, attention became scarce. When communication became abundant, trust became increasingly valuable. When digital reproduction became effortless, authenticity acquired a premium.

When creative content becomes effectively infinite, another scarcity may emerge:

The experience we know someone else had too.

That may become one of the defining cultural questions of the post-AI world.

The question will not be what we watch. There may eventually be more things to watch than a person could experience in a million lifetimes. It will not be what we listen to, because music could become effectively infinite. It may not even be what we create, because the boundary between imagination and production could become astonishingly thin.

The harder question is much more human.

What will we experience together?

Because in a world capable of generating a universe for every individual, the most valuable place may turn out to be the one place we still agree to meet.