The mark of an artwork is not simply connecting with the user, or to use a more blunt term, to engage with the user’s set of preferences such that it induces the necessary amount of dopamine to extract approval from its beholder. A true artwork, from the eyes of its beholder, would be seemingly impossible ex ante, and completely revolutionary after the fact. It would not match with one’s world-views. It would alter them in a way that is completely unrecognizable beforehand. Seeds of confusion and curiosity, sown by the artistic experience, will grow ever more. The beholder will become restless by its will to become one with the artwork, to adapt oneself, not the other way around, to the artwork so as to best be one with the world.

We keep asking whether we will be able to tell human-made and AI-made art apart. The more unsettling prospect is that our sensory habits will drift until the distinction stops mattering. Not because the two are metaphysically indistinguishable, but because the organs we once used to register difference—patience, attention, memory—atrophy from disuse. A famine of authenticity is coming, yes, but it will be experienced less as hunger than as satiety. We will feel full, continuously topped off by sequences that fit so snugly into our preferences that nothing has to be chewed.

This is the end of a long migration. The artistic experience has moved from a locus of suffering and reflection to a pipeline for pleasure. If the point of art is to deliver aesthetic enjoyment, why should its origin matter? A picture that resolves cleanly into the feeling I wanted—comfort, awe, melancholy—satisfies the brief regardless of whether it was coaxed out of a latent space or wrestled from weeks of human doubt. Once art is tasked primarily with delivering the right dose of affect, its provenance becomes a curiosity, not a criterion.

I learned this early, the wrong way. As a child I would stare at paintings and photographs so long that the air around them grew thin. By the time I could afford to travel, I felt I had already been everywhere. Standing before a cathedral I had studied in a book, or a coastline that had sat as my desktop background for years, I felt a quiet but unmistakable déjà vu. The shock that used to belong to encounter had already been spent in abstraction. When Neuralink or its cousins offer the sensation of touching grass, will I go outside less? The honest answer is not simply yes or no, but that the outside will shift from encounter to verification: a quick check that the simulation is accurate, then back to the stream.

Art has become a local event. By “local” I mean bounded by its own duration. It no longer inscribes itself; it does not migrate into a memory that gathers and thickens the self. It endures only while it is being perceived. The disappearance of “after” is not an accident but a design goal. Shorts have reshaped perception to prefer windows too small to store anything. Perception, in turn, has reshaped art to live only inside those windows. When the clip ends, it leaves no residue. There is no scar tissue, no sediment, no trace that memory can revisit and rework. The feed is a perfect solvent: everything that enters it exits weightless.

We have learned to measure art by engagement, which is to say by how efficiently it binds our preferences to content. The model learns what we like; the artist learns what the model likes that we will like; the platform optimizes the relay. But an artwork is not just a well-tuned hinge between desire and dopamine. The mark of the real thing—experienced from the side of the beholder—is that it feels impossible beforehand and inevitable afterward. It refuses to fit, then remakes the container. It arrives against our priors, not as their fulfillment, and then reorganizes them such that we cannot imagine the before. Confusion and curiosity are not defects here but the right kind of pain: the ache of new faculties forming.

This is why the provenance question cannot be reduced to aesthetics. A work that changes me is not just a delivery of pleasure; it is an instruction in how to perceive. It builds the very organ that will later recognize it. In an age that externalizes perception into algorithms, we risk outsourcing that construction entirely. The result is not merely kitsch or “low taste,” but a deeper loss: a sensory regime that cannot register what exceeds it. We will declare the excessive to be broken, the difficult to be inefficient, the slow to be obsolete. We will call it a hallucination when something real refuses to model neatly.

None of this requires a moral panic about tools. It requires a change in posture. If art is now local by default, then those who care about it must practice making it global again. Global here does not mean viral; it means capable of living after its duration, of being carried in memory, of continuing to pressure the self once the window is closed. That is a different optimization problem. It is not solved by better hooks or tighter edits. It demands forms that risk friction and invite refusal; forms that hold open a space where confusion can ripen into recognition.

There are modest, practical ways to do this. Refuse sequence when a single image can bear attention. Allow silence to remain unscored. Name things precisely rather than calling them “vibes.” Let titles misbehave when accuracy would sand away the grain. Most of all, resist the reflex to adapt the work to the user. Adapt the user to the work. Write, paint, compose in a way that demands the body reconfigure itself—posture, breathing, eye movement—to meet the work halfway. If a piece cannot survive a second look without another hit of novelty, it is not alive; it is just efficient.

The irony is that nothing prevents AI from participating in this demand. A machine can be used to construct ease or difficulty, to dissolve the self or thicken it. What distinguishes the two is not the tool but the telos. Are we calibrating to our preferences, or are we letting the work calibrate us? The former produces the average—comfortably smooth, immediately legible, forgettable the moment it ends. The latter produces an after: a pressure that persists, a quiet restlessness that sends us back not for more of the same, but to become different enough to meet it again.

If the senses are to be retrained, they must be given occasions to fail gracefully. Failure is not a bug but a pedagogy. The art that matters will be the art that first feels like a mistake and later like a map. It will make us leave the house not to confirm the simulation, but because the world has become legible at a resolution the stream cannot render. Touching grass will not be a wellness ritual or a verification step; it will be the only medium large enough to hold what the work has made us able to perceive.

Until then, the choice is simple and difficult. We can continue to inhale the local and let memory thin to the thickness of a swipe. Or we can seek the global: the works that resist us, confuse us, and then change the kind of creature that we are. The first path turns art into content and the beholder into a well-modeled preference. The second revives an older pact: the artwork remakes the world, and we remake ourselves to meet it.