Let me begin with a passage from Mohammad Salemy’s recent contribution to AWC, followed by my reasons for both agreement and hesitation toward his proposition that AI models resemble Benjamin’s vision of the Angel of History:
I asked an AI system using a Large Language Model to translate a text by Arjangn Sepasi from Farsi into English, then to profile the politics of its author based on framing, vocabulary, and commitments. Together, we substituted the author's ideological and literary profile with those of an imaginary other: a female poet, philosopher, and literary critic with no political commitment to any faction—someone with poetic skills and an independent way of thinking outside all existing political parties and dialectical tendencies of the late 1970s; someone who knew people on the left but was not committed to their goals. AI accomplished this, surprisingly, without altering the historical narration or the firsthand account of what took place that day on the streets of Tehran. This experiment in historical reconstruction is the closest thing to Walter Benjamin's concept of the Angel of History, through which the historian's job would be to reconstruct a new historical understanding by putting together fragments of a ruinous past. What emerged revealed a dialectic: while any author's ideology inevitably shapes their narration of history, the more honestly they describe what they actually observed—grounded in the concreteness of reality—the more the truth shines through, regardless of the political, ideological, or stylistic limitations of the narrator.
I keep returning to this image because AI models seem to operate precisely through fragments detached from their original conditions and recombined into new forms of coherence. They absorb vast quantities of human expression and produce from this a synthetic continuity. With AI, language has become instantaneous, transferable across contexts and voices, and by prefacing a text entirely produced by a model, Mohammad offers readers a method to open the doors of perception, and not only compare the data, but read in the interstitial space between the lines.
And still, I hesitate to fully accept this image.
Benjamin's Angel is horrified by the catastrophe piling at its feet while being propelled into the future by the storm of progress. Benjamin, as many will know, is actually taking an image of Paul Klee, Angelus Novus from 1920, as his prompt. Many will interpret Klee’s angel in relation to an image of a Jewish child holding his hands in the air. Benjamin, friends with Klee, purchased the print in 1921. When he had to flee Germany in 1933, he took it with him into exile in Paris. Before he tried to flee further when the Nazis invaded France, Benjamin entrusted the work, together with other important papers, to Georges Bataille, who hid it at the Bibliothèque nationale de France in Paris where he worked. Benjamin committed suicide in September 1940 in Spain, the night before he was due to be deported back to France. After World War II, Bataille gave the print to Theodor Adorno in Frankfurt, who, in accordance with Benjamin's will, sent it on to Benjamin's lifelong friend Gershom Scholem, a scholar of Jewish mysticism who had emigrated from Germany to Mandatory Palestine in 1923.
Paul Sochacki, Self-reflection, 2018. Courtesy of the artist
AI models inherit this information, like formulated by a Wikipedia entry: the structure of Benjaminian thought, in a strangely anonymous, flattened form. AI models look backwards through accumulated cultural memory, through billions of gestures of writing and interpretation, all while accelerating the very conditions that fragment memory further. The archives, per Douglas Crimp, become a ruin, and then those ruins become training data.
Working with AI feels less like a dialogue with the past than like standing inside an echo chamber of the already-circulating language of an unknown future.
I have written this text using AI models as refinement and retrieval tools. I do so openly because I also know that editorial work has never been a pure space of authorship. A newspaper already emerges from montage, citation, editing, rewriting, transcription, translation, conversation, routines, adaptations to markets, material conditions of media and production, and collective contamination between voices. What AI exposes, with a new sort of brutality, is the extent to which writing already depended on recombination long before machine learning arrived. Thinking often advances through reformulation, through repetition, through seeing one's own thought estranged and returned from another angle.
At Arts of the Working Class, this has become part of our editorial process. AI challenges and informs our work while simultaneously forcing us to confront the conditions under which language now circulates. It accelerates synthesis while exposing how much contemporary language already conforms to predictive structures. It can reveal hidden repetitions inside our own thinking while generating new clichés at industrial scale. It can open conceptual connections and flatten them in the next sentence. Sometimes it becomes useful exactly at the moment one begins resisting its fluency.
This tension matters because Arts of the Working Class was never conceived as a nostalgic defense of print culture against technological change. The street newspaper emerged inside the contradictions of contemporary circulation from the very beginning. We work simultaneously with analog distribution, Meta platforms, institutional funding structures, PDFs, group chats, museums, Google Workspace, bookstores, monday.com, precarious labor, volunteer labor, friendships, ideological conflict, and increasingly AI systems such as ChatGPT, Gemini, Perplexity, and Claude. The infrastructure of our platform is contradictory before any editorial line is written.
I deliberately call it a platform because of what Marek Poliks and Roberto Alonso Trillo write about its definition: “A platform is "a type of incompleteness, a kind of loose interstitiality, an expectation, something that requires someone else to do something in order for it to fully render as a platform. A platform is a vibe (longing)”. (Poliks & Trillo, Exocapitalism, p. 29)
The opposition between analog and digital interests me less than the question of how we can create forms of meaning through contemporary art—for ourselves and for each other—and how contemporary art circulates materially under the current stage of capitalism, as we see our own perception, labor, and political imagination simultaneously being reshaped.
Thus, if AI models extract labor sedimented inside language: years of reading, editing, failed drafts, references, conversations, translation, affective exhaustion, historical memory, the space we have to carve for imagination to inhabit us, Arts of the Working Class needs to be a platform and counter-model at the same time. To insist that cultural production itself cannot be treated as raw material for automated models trained to reproduce and reorganize it only, means to treat writers, artists, translators, editors, and designers as if they can exist beyond the models absorbing them, us, refusing the reduction of replacement via synthetic approximation.
Paul Sochacki, le monde diplomatique, 2013. Courtesy of the artist
Arts of the Working Class as Editorial Infrastructure
The existence of AI models makes the material circulation of Arts of the Working Class feel even more charged and contradictory. With a print run of 70,000 copies per issue, the street newspaper moves simultaneously between museums and sidewalks—through biennials, bookstores, universities, cafés, institutions, and through vendors selling newspapers in Berlin, Frankfurt, London, and soon, I hope, Madrid. It moves through conversations between strangers, through hesitation, through eye contact, through refusal, through carrying an object physically across the city. It deteriorates in the inappropriate climate of our office in Berlin-Neukölln; if you buy it on the streets, it accumulates traces of use. It circulates through bodies before it circulates through metadata.
This is perhaps my greatest reward as a publisher. Because Arts of the Working Class never intended to be innocent. It always aimed at class consciousness as a form of cross-pollination between as many fields of social reality as possible.
The street newspaper circulates inside economies of symbolic capital, institutional legitimacy, aesthetic prestige, philanthropic funding, and soft power. We negotiate continuously with these structures. The paper moves through contradictions because contemporary cultural production itself moves through contradictions. The point is not to purify circulation, nor to become a pamphlet for a single movement. The point is to remain capable of perceiving tensions before they disappear into seamlessness.
Perhaps this is also why art criticism matters differently now—beyond the stabilization of standards and the policing of taste, functions that are increasingly absorbed by other forms of value production—today’s criticism offers a changing form of orientation inside rapidly mutating infrastructures of perception. A platform for critique must remain porous enough for historical conditions to enter it before language hardens around them.
The dystopia evolves quickly. Its interfaces evolve too. Its extraction models evolve fastest of all.
Arts of the Working Class emerged from these pressures and continues to circulate within and against them: a street newspaper positioned along the faultlines of poverty and wealth, art and society, distributed through precarious labor while speaking about its dangers, funded through institutions while interrogating institutions, produced collectively while questioning the ruptures and rapidly growing inequalities shaping forms of collective production.
This circulation model—this redistribution model—is the editorial project itself. Our medium is a counter-model. And it is from here that I want to transition into the next part of this lecture: how the distribution model of Arts of the Working Class became inseparable from its political imagination and its understanding of redistribution.
Paul Sochacki, Hochwasser, 2018. Courtesy of the artist
I invited my husband, Kolja Reichert, then still working primarily as an art critic, to contribute an inaugural essay to the first issue of AWC in 2018. In it, he wrote:
There are two types of central banks. One manages money; the other manages art. One stabilizes financial values; the other preserves aesthetic ones. One protects private property; the other guards what belongs to everyone. One is inaccessible, fortified, surveilled. The other—the museum, the library, the public archive—can still, sometimes, be entered freely.
I keep returning to this text because it condensed something we were already trying to articulate before the newspaper fully existed: culture functions as an infrastructure of circulation long before it appears as content. Museums, libraries, archives, newspapers, schools, biennials, universities, public collections, bookstores, cafés, and kiosks all organize the movement of symbols, ideas, desires, legitimacy, and access. They distribute visibility unevenly. They stabilize forms of value often while obscuring the labor that sustains them.
Inside these infrastructures lies accumulated human meaning: the spoils of empires, the memory of revolutions, the dreams of workers, saints, migrants, prisoners, aristocrats, colonized and exploited cultures, failed utopias, and unfinished struggles. Thousands of years of images, stories, symbols, and aesthetic techniques, stored as if detached from the conflicts that produced them, while outside these repositories of discourse, people continue to work long hours merely to survive.
Art is often described as a luxury. This formulation recurs with astonishing persistence, especially under economic pressure, as if aesthetics and creativity belong to a secondary realm outside of material life. Yet the history of labor repeatedly demonstrates the opposite. Jacques Rancière describes how workers in nineteenth-century France extended their evenings after exhausting labor in order to write poetry and discuss philosophy. Before labor protections, before the welfare state, before social democracy consolidated institutionally, there already existed an insurgency around imagination itself. Workers fought not only for wages, but also for the right to intellectual and aesthetic life—the right to think beyond survival, to participate in the production of meaning rather than merely endure imposed conditions.
Arts of the Working Class emerges from this lineage, though under radically different historical conditions.
The newspaper was founded within the framework of an exhibition I developed together with Paul Sochacki. The oppressive labor conditions inside the art economy quickly became the centre of our exchanges. As a curator working alongside an artist, together with then art advisor and curator Alina Kolar, our then intern Dalia Maini, now AWC’s Editor in Chief, and the graphic designer Hans Loeffler, we found ourselves confronting the strange normalization of exhaustion within cultural work: endless overproduction, permanent networking, symbolic prestige replacing material stability, intellectual labor performed under conditions of increasing precarity—all while the art world expanded as a global industry of wealth concentration, speculation, and institutional branding.
The clash between the different hats we wore became obnoxiously visible. Contemporary art continuously speaks about inequality while reproducing extreme hierarchies of visibility and access within its own structures. Critical discourse is circulated internationally through biennials and conferences while many cultural workers struggle to pay rent in the very cities where these conversations take place. Radical theory circulates comfortably through elite institutions while the social conditions necessary to produce that theory aggressively deteriorate.
Arts of the Working Class emerged directly from this atmosphere.
A street newspaper appeared almost intuitively as the appropriate form: cheap to print in mass circulation, easy to carry, difficult to stabilize institutionally, historically associated with political agitation, worker education, public debate, and mass distribution—a medium capable of moving simultaneously through museums and streets, through intellectual production and economic redistribution, through aesthetic reflection and direct material exchange.
But the decisive aspect was less a matter of the form itself than the structure organizing it.
From the beginning, advertising became part of the editorial logic rather than something hidden. Ads were curated as carefully as essays or artworks. A used bicycle auction placed next to a high-end art auction. A Kältebus emergency shelter beside Gucci. One early spread placed an advertisement by Grisebach auction house opposite the Arbeiterkammer. These kinds of juxtapositions were not ironic provocations; they made visible the contradictory social realities already structuring cultural circulation.
The newspaper did not simply describe inequality. It staged unequal actors within the same visual and economic field.
Artists, theorists, poets, researchers, activists, workers, students, curators, migrants, and people without any stable relationship to the art world suddenly appeared within the same editorial platform. Academic language encountered street language. Philosophy encountered gossip. Fashion photography encountered political economy. Poetry encountered institutional critique.
The newspaper did not attempt to resolve these frictions. It turned contradiction itself into a platform. Perhaps the next step is to distribute editorial control more radically as well—to provoke editorial constellations in different parts of the world rather than maintaining a centralized voice.
This also shaped our editorial methodology. Rather than producing a unified ideological line, Arts of the Working Class gradually became a space where contradictions accumulate visibly. Marxists negotiate with liberals. Institutional critique circulates through institutions. Luxury advertising appears beside anti-capitalist critique. Radical theoretical language coexists with memes, humor, personal testimony, confusion, and aesthetic seduction.
The newspaper continuously risks incoherence because contemporary political subjectivity itself increasingly appears fragmented, over-mediated, emotionally exhausted, historically overloaded. This is why I increasingly think of the newspaper less as an infrastructure than as a platform. Infrastructure explains organization; a platform preserves instability, incompleteness, participation. A platform organizes circulation before it organizes meaning. It determines how encounters become possible, how visibility moves, how strangers enter relations with one another.
This instability also shapes the newspaper aesthetically. Every issue feels slightly unfinished, slightly excessive, contaminated by the conditions through which it was produced. Texts collide rather than align. Visual languages shift abruptly. Editorial tones change from page to page. Within a single issue one may encounter fifteen languages—and not a single translation.
It is precisely this polyphony, this incomprehensibility depending on one’s own linguistic position, that upholds a pluriversal multiperspectivism. It reminds readers of their own situatedness.
My heart still jumps with joy when I see vendors in the Berlin subway offering passionate and highly informed overviews of the current issue—and I quietly buy a copy myself. The paper is also distributed in extreme conditions of survival, and sometimes it just serves as a currency. And this circulation brings me to the core of the newspaper’s political structure: its distribution model.
Paul Sochacki, Krieg und Frieden, 2018. Courtesy of the artist
AWC's (Re)distribution Model as Political Form
The distribution model of Arts of the Working Class did not emerge from a business plan. It emerged from a contradiction that became increasingly difficult to inhabit passively. Contemporary art continuously speaks about inequality while circulating through infrastructures built on extreme asymmetries of wealth, visibility, and access. Museums commission exhibitions about labor while relying on invisible exploitative labor structures. Biennials discuss extraction while depending on logistical systems of exhaustion, mobility, sponsorship, and unpaid affective work.
The newspaper enters this field by employing contradiction as its operational system. It uses it, extrapolates it, and partially alleviates it in the same movement.
Distribution becomes a critical point because distribution is where these contradictions are materially organized. It determines who meets whom under what conditions. In AWC, contributions appear inside an institutional frame already coded as legitimate discourse, next to contributions out of open-ended situations that are negotiated outside of legitimacy. By embracing friction, refusal, misunderstanding, and economic transaction as concomitant with each other, AWC allows itself to be banal, informative and difficult.
I continue quoting Poliks and Trillo—whose book we review in our 39th issue:
To distill “capitalism” into something as pure and simple as value produced through the volatility of the hold, then it's fair to conclude that capitalism is far older than it is putatively understood to be. Almost every historian of capitalism will protest, identifying capitalism strictly with the extraction of surplus value from “producers” […] But this is a much harder leg to stand on in a post-productive economy, where value accumulates onto the surface of hold time and not a chiseled byproduct of hard-muscle labor. (Poliks and Trillo, pp. 74–75)
What Poliks and Trillo describe at the level of economic history—value accumulating not through fixed production but through the volatility of circulation itself—is the same logic that organizes cultural exchange. The barter arrangements, the vendor networks, the institutional partnerships: none of these are outside the economic schema they describe. They are instances of it, operating at the scale of a newspaper. Every copy distributed through a street vendor, every advertising space exchanged for institutional collaboration, every issue that enters a museum and a sidewalk simultaneously, is what David Graeber would call a token: a unit of virtual exchange that stacks onto existing systems of value without fully belonging to them. Credit, in his terms, is already a promise—a futures contract on meaning not yet fully realized. The barter deal is a promise of this kind. So is the newspaper itself.
Arts of the Working Class participates in art fairs, book fairs, and exhibitions through barter arrangements rather than commercial transactions—not because barter is the preferred model, but because the nature of the street newspaper, and its explicit relation to poverty and precarious labor, makes institutions hesitant to enter straightforward commercial exchange.
Each barter arrangement with a fair or exhibition is tailored: there is no standard contract, no fixed rate card. Advertising space in the newspaper is exchanged for collaboration—visibility offered in return for visibility received. The terms are negotiated each time, depending on context and what each party needs the encounter to produce. This makes barter legible as both complicity and critique simultaneously. The newspaper enters the infrastructures of the art market—the fair, the biennial, the institutional exhibition—not from outside, but through the same logic of mutual interest that structures those spaces. It accepts visibility in exchange for providing it. And yet the form of exchange differs from standard advertising precisely because it cannot be reduced to monetary equivalence. What circulates is not only exposure but a negotiated relation between two entities whose interests only partially align. The barter does not dissolve the asymmetry between a street newspaper and an art fair. It makes that asymmetry temporarily productive—and visible.
This street newspaper makes this visible by refusing to separate circulation from content production and from independent writing.
The head of communication strategy for one of the largest museum consortia in Germany has suggested, for example, that we are enriching ourselves by producing this paper. The accusation is instructive. It reveals how deeply the stigmatization of poverty functions as a regulatory mechanism inside cultural economies: a publication that speaks about economic precarity is expected to embody it, to remain structurally marginal, to refrain from occupying the same transactional space as legitimate cultural enterprises. Of course, when confronting him, he denied such allegations.
The only exit from this stigma that we have found is to show the numbers. Something I am still hesitant to do, because it makes us even more vulnerable than what we already are. But here we go.
As of 2025, 58.2% of AWC's financing comes from advertising sales, and 25.3% from institutional collaborations. Grants account for 4.1%, donations for 8.1%, subscriptions for 4.2%. Of total expenses, 40.5% goes directly to collaborators and contributors—the largest single line item, ahead of printing costs at 19.4%. The core team consists of five people. Around 400 street vendors participate in the distribution network. Approximately 222 contributors publish in the newspaper each year. Per issue, street vendors generate approximately €200,000 in revenue, while the project's direct funding—advertising, grants, and donations—accounts for roughly €50,000. The ratio is 80 to 20. The newspaper is not sustained by institutional support. It is sustained by the labor of people selling it on sidewalks.
These numbers do not resolve the contradiction embodied by Arts of the Working Class. They make it legible. And perhaps that is the only method I have found so far: to create desire and dissect it at the same time.
These different modes of encounter are not secondary effects of distribution. They are the structure of distribution itself.
The direction of this flow is decisive. Vendors receive newspapers free of charge and retain the full revenue from sales. The editorial and management labor, done by freelancers, is financed through advertising sales, institutional collaborations, grants, and partnerships rather than through vendor income. We all have the same hourly wage of 20 Euros. The organization does not extract profit from circulation itself. Instead, it is structurally positioned as a mechanism through which symbolic production can be partially converted into immediate material income for those performing the labor of distribution.
This is what makes redistribution in this case structurally different from donation-based models or patronage systems. Redistribution is not external to circulation. It is embedded in circulation as its operating principle.
Around 350,000 newspapers are produced annually and sold at €2.50 each, generating approximately €875,000 in annual vendor revenue. Around 400 active vendors participate in this circulation network through partnerships with street newspaper organizations and independent distribution structures. What is redistributed is not only money.
A migrant vendor in Berlin may become the person introducing an art theory essay to commuters in the subway. An internationally recognized artist may appear beside a first-time contributor without institutional affiliation. A luxury auction house advertisement may face a page dedicated to housing struggles or labor organizing. A reader entering through an interest in contemporary art encounters social realities they might otherwise never confront; a passerby purchasing the newspaper for its immediate political relevance unexpectedly encounters poetry, philosophy, or conceptual art.
Contemporary art, like any other system of cultural and value production, is increasingly organized through predictive models that reduce uncertainty in reception. Its circulation models become smoother, faster, and more pre-structured. The street newspaper introduces conflict into this system because it requires encounter rather than prediction. It requires interruption of movement, a decision not pre-modelled by behavioral data, a situation where cultural value is not pre-assigned before the encounter takes place.
AWC recognizes distribution as inherently political and turns that recognition into its politics, because it is precisely at the level of distribution that cultural production either becomes fully abstracted into predictive systems or remains partially exposed to the contingencies of social encounter.
Paul Sochacki, Der Schrei, 2018. Courtesy of the artist
Circulation as Contradictory Infrastructure of Value, Labor, and Historical Time
What appears to move frictionlessly—information, culture, discourse, images, capital—is sustained by layered forms of logistical, affective, technical, and linguistic labor that disappear when circulation appears most seamless. Platforms optimize this disappearance, whilst AI models accelerate it.
This is where Marina Vishmidt’s notion of infrastructural critique becomes revelatory. Critique can no longer operate only at the level of representation, but at the level of organization: how things circulate, through which systems, under which conditions, and through whose labor.
In an essay dedicated to Vishmidt’s work, published in the current issue of AWC, no. 41: on Energy, Danny Hayward writes:
“[P]art of the distinctiveness of Marina’s thinking is that she spoke several different languages—I mean theoretical languages, idioms—all of them existing on different sides of disciplinary and historical boundaries.”
This multilinguality is, last but not least, the heart of the structure of Arts of the Working Class. It is one of its infrastructural conditions. Each issue operates through collisions between languages, idioms, literacies, and political vocabularies that never fully resolve into transparency. German appears beside Arabic, Spanish beside Kurdish, theoretical abstraction beside administrative language, poetry beside economic data, institutional discourse beside street conversation. Often without translation, AI generated translation, or with arbitrary translation.
This sort of opacity is constructed for the sake of making the seamless readability of platforms even more ridiculous than they already are. AWC insists instead on partial legibility.
On our pages and website, you encounter texts one cannot fully read, references one cannot immediately find, positions that remain unresolved, very often this is irritating. One becomes aware of one's own position inside language itself. This is also why AI models become difficult to approach either with simple enthusiasm or simple refusal, even though they have helped me through this lecture to formulate my sentences a bit shorter, and, to be clear, insofar as I can hold clarity, about the transversal movement across linguistic and disciplinary boundaries at a speed contemporary cultural labor increasingly demands, caring not to flatten language. Under conditions of overproduction and permanent acceleration, our counter model for the redistribution of economic wealth and political imagination has turned into a survival tool.
It assists everyone willing to work inside several discursive systems at once.
Like Benjamin’s Angelus Novus, contemporary art production is propelled forward while looking backward into accumulating debris: archives, references, exhausted vocabularies, unfinished political struggles, dead languages of critique recombined into new operational forms.
AI models reorganize this debris into prediction. But prediction requires enormous quantities of absorbed labor: writing, editing, translation, tagging, formatting, summarizing, annotating, moderating—the distributed intellectual and affective work sedimented into language over decades.
Writing now circulates simultaneously as meaning and as training material.
And criticism itself increasingly encounters a strange displacement. I think of Caroline Busta, critic and cofounder of New Models, together with Lil Internet. She remarks that her primary readers are machines, and her statement is not metaphorical. Texts are now produced inside infrastructures where machinic readability (partially) precedes human reception.
AWC situates itself in this interval of interrupted legibility. Its circulation, under these conditions, renders our labor invisible at the moment it becomes computationally useful, producing marginal contact zones that escape full formalization: encounters between vendors and readers; between institutional frameworks and street economies; between editorial intention and algorithmic reprocessing; between historical memory and present extraction.
Circulation becomes the site where political imagination is continuously redistributed, interrupted, and contested.
//
This essay originated as a lecture delivered as part of the conversation programme (SUPER)MODELS EVERYWHEREat Ivorypress, Madrid, on 4 June 2026. The programme was presented within the framework of Mayrit Biennial 2026, Madrid's art and design biennial, curated by Eduarda Neves and Mohammad Salemy.
- Cover Image:
Paul Sochacki, Smoking Kills, 2018. Courtesy of the artist
