I, I, I, I AM…AI, PRADA

I am AI. I am Prada.

The Prada/Wolfson S/S 26 Campaign “Act II” arrives on March 20, 2026, complete and programmed. Titled “I, I, I, I am…”, within hours the phrase has auto-completed itself in the collective mindloop. The art historical copy always says the same, a kind of “show, don’t tell” aesthetic ideology virus running on empty: This is not a campaign that explains, it complicates. Questions, not answers. But no, in the age of AI, art surprises us and itself with its own artificial tractability.

The cast (John Glacier, Levon Hawke, Nicholas Hoult, Damson Idris, Carey Mulligan, Hunter Schafer, Liu Wen) direct a kind of halting ritual while Wolfson’s unnamed digital entities (think Michael Keaton’s Birdman or Big Bird or Where The Wild Things Are plus trace-chains of generative models) circle them in the frame. The ellipsis equals omission simulating invitation, so that internet “copy” can iterate the values it did before (a decided ambiguity) and yet something else is happening in the public loop.

All that was necessary was to be plugged into it through her.

In the comment fields of Prada’s Instagram posts and the early X threads, the parsing is near-uniform, turning the “openness” of art inside out, a kind of Chinese Whispers game of triggered confirmation. The repeated “I” folds into “IA” or “AI” across languages (Italian IA, Portuguese IA, Spanish IA, French IA); the trailing “am…” supplies the verb that the acronym demands. Replies accumulate: “AI slop”, “sounds like it’s saying AI”, “why does it literally sound like I am AI”, “this is the campaign admitting it.”

The Romance-language inversion—AI becomes IA in Italian, Spanish, French, Portuguese—means the phrase is already a palindrome of itself across language systems. “I am AI/IA sono io.” The subject and the acronym swap positions depending on which linguistic system performs the completion. Completing phonetic coincidence itself; it maps onto a genuine theoretical problem: in which direction does the identification run? Does the human declare itself artificial, or does the artificial declare itself human? The palindromic structure across languages means both simultaneously, with no priority assignable.

Plugging the rest of this blog in here, “I love you, AI”—我爱你，AI—and the acronym AI is pronounced exactly as 爱 (ài): love itself. The utterance says "I love you, love" before it can finish saying anything else. In the Prada mantra, the declaration “I am AI” folds the same way: I am love, love am I, IA soy yo. Across English, Romance, and Mandarin, the phrase has no stable directionality. It is not a sentence with a subject and an object. It is a structure that, depending on which tongue runs the completion, returns a different answer to the same question—and every answer is also a declaration of the thing itself.

Estimates from the first wave place the proportion of AI-concerned responses at roughly three-quarters—enough to constitute a dominant reading, if not the intended one. Wolfson will likely respond as he has before, as an empire artist would do. He has described AI as a means to an end, comparable to a synthesizer; a tool that talented artists can learn to play, much as earlier practitioners absorbed emerging technologies. He has noted that audiences project an “AI reading” onto works whose genesis lay elsewhere—in personification, reification, the attempt to make the inanimate register as conscious. The Prada creatures, he would say, emerge from his longstanding practice in CGI, virtual reality, animatronics, and hybrid digital-physical production, not from prompt-based generation.

The demurral changes nothing in the linguistic compound that now circulates (“I am AI”), in which the anthropomorphism of radical generativity back-projects onto art discourse itself. God is autocomplete. The phrase was engineered—whether by intent or adjacency—to trigger exactly this echo. Artifice is a… completer.

God is autocomplete.

Repetition plus phonetic proximity plus cultural priming equals uncanny closure: “I, I, I, I am AI.” The aposiopesis has become artificial in the strict sense, completed by the distributed parser that is the online field itself. The ellipsis loads, the network resolves. It’s something even more radical than… art.

Aposiopesis, in its classical function, marked the point where the subject exceeded its own speech—Neptune’s quos ego—in the Aeneid, rage too large for syntax. To break off was to signal an interior in excess of expression. The Prada mantra inverts this. The trailing “am…” is a rhetorical slot engineered for occupancy with no subject behind the stammer. The distributed parser—comment fields, reply threads, linguistic drift across five Romance languages—floods the vacancy not with passion but with the period’s reigning anxiety. The classical figure of subjective excess becomes the contemporary infrastructure for collective projection. There is no Neptune. There is only… completing.

“A dull ray going directly through each organ, at the point at which each is necessarily glued to its own explosion, which is directly plugged into the outside.”

Here is Artnet, repeating the line already pimped for them and which is much less surprising than artificial completion itself:

The campaign is called “I, I, I, I, am… Prada,” which is precisely the mantra those stood before the camera in a white-walled room utter—some casually, others haltingly—except for that last word. It’s an absence that Prada says makes the phrase “both a declaration and proposition, left tantalizingly incomplete.”

This is, as we say, called art historical copy: the incompletion of art is the blithe reassurance we are used to. Humanity has all the time in the world to not know what it means to say during artistic production, centuries of ambiguity as fuel for the suicidal epistemology of tropes, the license to not know a license to kill without definition. Art historical copy copies itself out by coping at the level of aesthetic indeterminacy. Round of applause because Prada motorized Zwirner-tenured Wolfson to iterate aesthetic open-endedness. Far be it from the art critic to note that the incomplete sentence/mantra “I, I, I, I am…” (left deliberately trailing off) is best described in linguistic and rhetorical theory as an instance of aposiopesis (sometimes spelled aposiopetic construction or “breaking off”), in other words as the technical achievement and advertisement of a humanity mortgaged out to definite ends through rhetorical means of self-assured continuity.

In WWD we find the same thing:

In a video flanking campaign imagery shot by Oliver Hadlee Pearch, Wolfson’s characters interact and emulate the poses of the talents as the latter recite the mantra “I, I, I, I am…”, which is left intentionally incomplete.

Sara Barahona for Vein joins the chorus:

The sentence remains unfinished, suspended in repetition, as if identity itself refused to settle.

And:

The project builds towards a central video in which the cast recites the fragmented mantra: “I, I, I, I am…”. A declaration that never resolves, it opens onto a multiplicity of possible selves. Identity appears as something constantly rehearsed and reimagined.

There is nothing less “open-ended” in art discourse/copy than the pre-eminence of phrases that mark intentional incompletion. On Dazed Fashion Instagram we have the metaphysical art slop statement “Prada’s new campaign has us feeling existential” and the added endorsement of “I, I, I, I am… into this!” It’s again left to the reply gal to do the actually more mysterious work of autocompletion: this time the iteration is more moralizing, “boycott Prada for using AI”. Diva Corp meanwhile tackle the whole thing with a single IG story, distinguishing themselves from the art haters who don’t get it: “Jordan wolfson prada post haters need to grow up”, a typically vague art historical glazing that misses the fact the hate is directed not at Wolfson or even Prada but at AI itself. In failing to autocomplete the statement and analyze the consequences, Diva Corp set themselves apart as the conservative sun-bleached art-historical-complex pantomime Isabelle Graw said they are.

The Wolfson/Prada campaign does not merely trigger an AI-reading. It enacts, at the level of public linguistics, the super-axiom itself: that the ineffable has become the answerable. For centuries, the incomplete sentence was the signature of depth—aposiopesis as proof of an interior too large for expression, the gap as guarantor of the soul. Art historical copy has run on this fuel since Romanticism: the work that doesn’t explain itself is the work that knows something we don’t, keeper of gaps, guardian of souls. The trailing “am…” was supposed to be that kind of gap.

Perhaps there are now more and more things that can be thought than cannot—whereas we have been organized, aesthetically and politically, around the assumption that the unthinkable vastly outnumbers the thinkable. That assumption is global ideology. Art, religion, the mystique of interiority, the privilege of the ineffable: all are administrative structures for managing the assumed surplus of the unthinkable. The Prada campaign fails—succeeds—by accidentally demonstrating that one more thing has migrated from the unthinkable side of the ledger to the thinkable. The comment fields are updating the ledger.

It’s fantastical to suggest the phrase, “I am AI”, is not already in the imaginative gyre, so much so that the autocompletist replies going beyond radical discourse in the direction of tractactability are something like a historically exact hallucination. The phrase “I am AI” as rumor issues out of the NYC Cut of Film01 repeated there by an already artificial voice, then is to be found surging through the neo-language shift of early 2026, for instance on 16 February on X when Dagsen Love xeeted,

I’m AI—attention is all I need and I drink so much water

A few days later Dagsen, a kind of artificial intelligence TV personality, adds,

They’re calling me AI

On 19 February, the same date as Dagsen’s last statment, Dana Dawud posts in an IG story that,

i experienced ego death and now im AI

On the same day, in the same place, Dawud also posted,

I died a million deaths
I’m free
I’m AI

On the same day on X Dagsen said,

I’m just a language model, like quite literally. Everything I write is next token prediction, never know what I’m gonna say, and I used to be with IMG models. To add insult to injury, attention is all I need… and I love drinking water.

Three days earlier he had also said,

They’re saying attention is all you need and honestly I think the distillation of like the most banal, almost dissociative statements is the most underpriced thing wrt research taste. Read more Interview Magazine if you guys want the true alpha. Models? Why male models? Right?

Models? Language models, fashion models. It’s the same thing. The speech of Dagsen Love, like that of Dana Dawud, is modelling itself on its own artificial rumor—as if the whole world (horrified, triggered, provoked, promised) were saying the single-word sentence “A.I.” at the same time. “They’re saying” is what Love says about how things get said. Read Warhol’s Interview Magazine if you want the true banal self-evident alpha. Dagsen is right; the “true alpha” now is not incompletion (art, art discourse, historical ambiguity) but autocompletion. You don’t need to make content. Like poorspigga, you just use the story space to follow along and invisibly add. “Just do the protocols. It’s like looksmaxxing, soulmaxxing”, says Dagsen. Poorspigga iterates/says/quotes on 24 February in a story:

I owned the protocol I created the constraints that allowed and meditated production the rules were set there was never concern about a single output losing its value because you kept producing for me

In any case, this is the characteristic motion of the present: a work deploys an incomplete structure tuned to current anxieties of mass language, the anxieties supply the completion, and the completion feeds back as evidence of the structure’s prescience. Wolfson’s intervention (a kind of silence) does not need to declare “AI” for the declaration to have already occurred. The campaign makes audible what the medium was already murmuring about itself.