The Signature and the Shadow
From the Turing Test to the Picasso Test
When we walk through a museum, we often know before we read the label who painted what. A Hopper evokes loneliness and light. A Kandinsky vibrates with spiritual geometry. A Magritte whispers riddles with clean surrealism. The mark of a mature artist, we might say, is recognizability. A signature.
But here’s the irony: what is most recognizable is also most replicable.
Midjourney and other AI tools can now churn out “new” works in the style of Van Gogh, Basquiat, Hokusai, or Rembrandt with startling fidelity.
The very thing that made these artists great is now the very thing that renders them banal. We may be entering a strange new phase in aesthetics: the more distinctive your style, the more easily an AI can mimic you.
If the Turing Test tests AI’s ability to pass as human, let the Picasso Test signify an artist’s ability to become AI-replicable.
You’ve passed the Picasso Test if an AI can forge your style.
Imitation as Training
In 11th grade, I took AP Language and Composition. One of the strangest, most formative assignments we were given was to hand-copy essays by great writers. I transcribed Emerson’s Self-Reliance, Thoreau’s Civil Disobedience, Dr. King’s Letter from a Birmingham Jail. I wrote out JFK’s and FDR’s speeches. I copied Coleridge and De Quincey in longhand.
At the time, I was annoyed by the drudgery. There’s a difference between watching the Karate Kid, and actually having to do the menial labor so glorified in the film. My hand hurt from hours of handwriting. It felt archaic, even monastic.
But over time, something shifted. By copying the rhythms of others, I started to hear my own. I began to notice choice: in sentence length, in diction, in tone. Style was no longer invisible, it was architecture. And with that came freedom, the emergence of my own style. That class single-handedly taught me how to write.
In hindsight, it was the same training method used by the greats of antiquity. Cicero learned rhetoric by imitating Demosthenes. Medieval yeshiva students memorize Talmudic arguments in sing-song. All of us, human or machine, begin with imitation. The question is what comes after.
As Picasso himself says, no doubt ripping someone else, “Good artists copy, great artists steal.”
The Anxiety of Influence
Harold Bloom, in The Anxiety of Influence, argues that great artists live with a deep insecurity: they know they are not as great as some of the giants who influenced them. Or sometimes they stand in the shadow of their contemporaries. There’s a powerful scene in Amadeus, where Salieri recognizes he’ll never hold a candle to Mozart.
Great poets, says Bloom, are shaped by the voices they admire, even when they want to reject them.
For Bloom, the strongest poets are not those who deny influence, nor those who merely echo it, but those who wrestle it into something unrecognizable. They “misread” their precursors, or rather transform them so profoundly that the lineage seems reversed; so that Blake seems to have authored Milton, or Whitman seems to stand upstream of Emerson. Or, to take a different example, the rabbis express their own anxiety of influence when they imagine Moses, lost, in the back row of Rabbi Akiva’s classroom.
In traditional Jewish learning, the stories we tell about biblical figures are not drawn from the plain text, but from a Midrashic overlay so powerful that it eclipses the source. Ask a child raised in a traditional Jewish home about Abraham, and you’ll hear about him smashing idols, reasoning about the sun and moon as false gods, discovering a Prime Mover through philosophical inquiry. None of this appears in the Torah itself. The biblical Abraham simply hears a call and follows it, without prelude.
Often, Rashi is more memorable than the verse on which he comments. We are so steeped in commentary that we forget what was commentary. The text becomes inseparable from its transmission.
AI and the Inversion of Anxiety
So how does this all relate to AI?
There is a new kind of anxiety of influence forming, not just our anxiety about being influenced by the past, but our anxiety about being imitated by the future.
Artists have always feared irrelevance, but now they fear redundancy.
When an AI can write in your voice, or even, write better in your own voice than you can, what happens to the meaning of authorship?
The imitation machine becomes a mirror, and we begin to ask not only: “Who influenced me?” but “Who will I be mistaken for?”
Roland Barthes famously pronounced the death of the author in 1968. Though foresightful and descriptively poignant, Barthes’ postmodern approach has also been a headwind for Humanities departments. If the author is dead—if meaning is whatever you want it to be—why be an English major? Why read Shakespeare? Humanities departments have gradually turned from places of literary reading to places of “literary theory.”
AI takes Barthes’s destruction of artistic intent and pushes it even further: now the author is not just dead; he may never have lived. When a convincing voice is produced by statistical inference and pattern recognition, the authority of experience collapses. The copy precedes the original. The output becomes orphaned from biography.
Benjamin and the Aura of the Inimitable
Walter Benjamin, writing in the 1930s, diagnosed the loss of “aura” in mechanically reproduced art. The original, he said, has a presence, a here-ness, a now-ness, that cannot be duplicated. A painting on a wall is different from a print. A performance is different from a recording. Reproduction changes not just form, but meaning.
AI adds a new layer. It doesn’t just reproduce objects, it simulates creators. And yet even here, there is a paradox: the aura may not disappear, but shift.
What is the aura in an age not just of mechanical reproduction, but generative AI?
From Copy to Creation
We used to imitate great writers to find ourselves.
Now, machines can and will imitate us, even before we’ve figured out who we are. This is either terrifying or a creative challenge or both.
Take Stephen Curry in basketball. Everyone wants to shoot like Steph. That makes him a model to emulate and a target to replicate. Style invites both reverence and reproduction.
The same goes for writing. We know Joan Didion when we read her. We know Orwell. Baldwin. Murakami. Their sentence rhythms and rhetorical tics are so recognizable they’ve become almost genres. AI can mimic them, but can it strongly misread them, as Bloom would say? Can it misunderstand them, and thus produce its own originality?
Time will tell.
The Co-Authoring Horizon
In the future, great style may come not from evading imitation, but from collaborating with it.
Artists will learn to use the machine without becoming the machine. They will develop styles that are trained against, through, and alongside AI.
And if influence becomes two-way, if we are shaped by AI, and it by us, then the challenge will not be to escape the loop, but to deepen it.
To write in a voice that is both machine-readable and human-irreducible.
So perhaps we are not witnessing the death of the author, but the birth of the author prompt, a set of constraints and styles so distinctive that others can simulate you, and never replace you.
You are recognizable. You are replicable. But you are not exhaustible.
That, perhaps, is the new goal of style.
- Zohar Atkins, Lightning
P.S.—If you are a professor, teacher, or home school parent interested in early access to Virgil, an Intellectual Companion, trained on 4,000+ Great Books, and a Socratic Pedagogy, reach out! If you have a pitch for an essay for Second Voice, email us: hello@lightninginspiration.com