WHAT IS THE FUTURE OF THE IMAGE IN THE AGE OF CLIMATE CHANGE AND ARTIFICIAL INTELLIGENCE?
Hito Steyerl is one of the most celebrated artists of our time. Her work, both as an artist and a writer, has consistently challenged the political boundaries between art and technology. In this new collection of groundbreaking essays, she explores how AI, the use of large language models and the algorithmic creation of imagery transform our understanding of the world. She argues that such practices cannot be divorced from the economic and political conditions of the times.
Medium Hot is a collection of scintillating meditations on the limits of art and technol the essential handbook for the present conjuncture. The pieces here probe the manufacture and distribution of images in the age of AI and climate change. She asks whether art can be made not only by machines but for machines. She argues against the production of images that heat up the planet, disfranchise workers and fuel the arms trade, and questions whether such creations can even be called art.
In an era of such rapid change, Steyerl does vital work investigating whether machine learning will infiltrate every aspect of our lives and what that means for the future.
Hito Steyerl (sometimes spelled Štajerl) is a German filmmaker, visual artist, and author in the field of essayist documentary video. Her principal topics of interest are media, technology, and the global circulation of images. Steyerl holds a Ph.D in Philosophy from the Academy of Fine Arts Vienna. She is currently a professor of New Media Art at the Berlin University of the Arts.
“Las organizaciones mediáticas más poderosas del siglo XXI serán térmicas. La circulación de imágenes, sonidos, videos y textos dependerá de un régimen masivo de calentamiento y enfriamiento. Los datos y las redes, al igual que las personas que estos conectan, serán cada vez más frágiles.
Si hay demasiado calor o demasiado frío, las plataformas colapsarán. Las infraestructuras digitales -centros de datos, intercambios de red y cables de fibra óptica- drenarán la energía del planeta para crear un entorno térmico estable, no para las personas, sino para la información.”
el mundo adaptándose a la IA, el basilisco de Roko, los intereses militares detrás de todo esto… Ahora más que nunca la inteligencia artificial está más presente que nunca (quién no conoce a alguien que utilice ChatGPT, por ejemplo); Hito Steyerl nos plantea varios escenarios relacionados con la misma, dejando varias preguntas y reflexiones muy interesantes sobre cuál es el rumbo de estos propósitos en la cultura, el arte y las guerras.
This was a strange read. I'm kind of in the middle ground of knowing-things / not-knowing-things about AI and other digital buzz topics so some of it went over my head (or felt dubious, but I didn't have enough info to argue why). When I was at a point of struggle-incomprehension I just pretended this book was a weird poem. I think these ideas will probably come back to me in the future though!
Steyerl goes into a little more detail about topics that have been heavily discussed already, like AI and cloud infrastructure's impact on electricity and water flows. It was interesting seeing direct examples of how this impacts local populations, like in Kosovo. Horrifying hearing about how ML training tasks/microwork are offloaded to displaced persons, and how some AI companies are faking their ML to just use these workers on the user-facing end.
I liked the discussion of how and why images are produced using AI. I can see how creative industries might be used to sort of humanize this tech, or distract from its primarily military applications. Steyerl also talked about how censorship impacts image-generation output, and how AI's artistic literacy might be seen as a stepping stone towards AGI (through "common aesthetic sense"). That last part gagged me a little I cannot lie. IS art something that does that for humans? Actually, I think language does, and art is just a more arcane and socially-loaded language... Much to consider above my level of judgement. What the hell.
I did not care about the NFT discourse or crypto discourse, and Steyerl was sometimes really unfunny and nonsensical to me, but I think that is a problem with tech/art people and millennials generally hahah. I did enjoy this book a lot.
EDIT: She also mentioned some philosopher's very insane idea of Capital as a higher-level entity - something about the large amt.s of energy collection/transformation thru capital - VERY crazy but let me ponder this hmm
Me convertí en una atenta lectora de Hito Steyerl. En Medium Hot logra un buen balance entre lo técnico, lo especulativo, el humor ácido, y el arte. Para mí, sus libros son actuales, fáciles de abordar y necesarios.
Este libro toma como aspecto central de qué manera el auge de la inteligencia artificial, los datos y los sistemas automatizados están reconfigurando no solo la producción cultural, sino también el trabajo y hasta la noción de sentido común. Las imágenes ya no representan: extraen, dañan, acumulan y se autocompletan. Para ella, en este contexto el arte funciona como laboratorio, pantalla de humo y campo de batalla. Todo al mismo tiempo.
Uno de los puntos más fuertes está en el capítulo final, una suerte de fábula sobre futuros posibles del mundo del arte, que combina sátira, ciencia ficción, tests de la revista Cosmopolitan y una fuerte crítica institucional. Es un cierre brillante con una tremenda potencia conceptual.
Creo que este libro logra ser muchas cosas a la vez: ensayo político, manifiesto cultural, diario. Es para aprender, subrayar y aprender mucho. Ideal para intentar entender nuestro presente digital y algorítmico.
I have to admit that I am not especially interested in what Hito Steyerl has been writing lately. The reason is simple: I do not feel that what she has produced over the past fifteen years has moved much beyond the same recurring concerns—critiques of the art world, discussions of the image and technology, arguments that often feel repetitive, wrapped each time in a slightly different comic sensibility. The Wretched of the Screen remains, in my view, the best book (essay collection) Steyerl has written, and what has come after it has felt rather ambivalent. This time, in Medium Hot, a collection of essays on AI and its relation to climate catastrophe, neoreactionary politics, and, as usual, the art world, I still do not find the excitement I once got from the treatise “In Defence of the Poor Image”. My assumption is that Steyerl’s pre-social-media-caption style of writing — dense with jargon, punchlines, pop-cultural references, and dry humour—simply no longer produces much of a reaction in me, so reading Medium Hot feels rather bland. Meanwhile, if I want to encounter ideas that are sharp, important, and worth dissecting on the topics Steyerl addresses here, I can find far more of them in contemporary philosophy of technology or media theory. One of the strongest essays in the book, in my opinion, is “Burnt-Out Images”, and even that is because Steyerl engages with a range of ideas from Bernard Stiegler's negentropy, to the grand master of STS, Donald MacKenzie. In any case, regardless of my reservations about her style, or the fact that the substance of the ideas here does not feel especially informative or substantial to me, I fully agree with Steyerl on one point: the AI race, or the teleology driving towards AGI, has so far produced almost nothing that is genuinely transformative enough to reorder the world for everyone. The AI race functions largely as a race for its own sake.
Steyerl's previous output often highlights unexpected connections between art, images, technology, geopolitics, and capitalism. Generative AI sits right at the crossroads of those topics, so I was very excited for this.
Where Steyerl's previous book of essays (Duty Free Art: Art in the Age of Planetary Civil War) was a rollercoaster ride of disjointed topics, the narrower focus makes this slightly less thrilling but a lot more readable. Given how much AI discourse there's been everywhere, some ideas might be familiar: - AI gives statistical averages not creativity. - Training is based on the internet and not on reality. - Gen AI images are fundamentally less useful than real images. - AI companies are rentier capitalists selling our stolen data back to us. But it goes far beyond that: the terrifying views of e/acc fanatics, and the insane story of Steve Bannon's involvement in an experiment called Biosphere 2. There's also a mostly-convincing call for DAOs to explain why they can't just be good old-fashioned co-operatives with an explicit social contract. The standout chapter was We Didn't Start The Fire, highlighting how conflict can create big tech's beloved inputs of cheap energy (conditions in Kosovo driving a grid-destabilising amount of crypto mining) and cheap labour (refugee microworkers labelling data for training sets, or reviewing horrific illegal content).
At times, I struggled to tell how serious she was being. I have no idea what the last chapter (21 Art Worlds) was trying to do.
Jeg har i løbet af de sidste par måneder forsøgt aktivt at forholde mig til, hvad såkalt GAI’s indtog på verdensscenen betyder for vores kommunikation, vores videnskabelse og ikke mindst for kunst og billeddannelse…. Det har ført til øjeblikke af skiftevis fuldstændig afmagt og apati (tekst-er-dødt-følelser), vrede, forvirring og afsky…
Hito Steyerl forsøger som en af de første at sætte, jeg har læst at sige noget reelt intelligent om fremkomsten af GAI. Om overgangen fra kausalitetsbetonet Newtoniask fotodannelse til det propabilitetstermodynamiske shitshow, der er AI-generede billeder… om overgangen til det gennemsnitlige, om den infrastruktur af mikroarbejdere i det globale syd (og ikke mindst i konfliktzoner, hvor elregningen ikke skal betales) som kapløbet mod AI-entropi baserer sig på.., om forbindelsen mellem det hybris-prometheus neo-fascistiske fundament som tech-giganterne bygger på og den fortsatte opvarmning af kloden.
Det gør den tyske verdenskunstner klogt. Bogen leverer et foreløbigt omend ufærdigt bud på et sprog, som vi kan benytte til at tale tale om GAI uden bare at gentage OpenAI’s egne reklamefloskler om en superfed post-arbejde (men ekstremt ulige) fremtid, hvor du aldrig overbooker din kalender!
Reading this in 2025, the essays themselves still seem for the most part pertinent (and sometimes prescient), though some of the sample images no longer represent the capabilities of generative AI image diffusion models. (I skipped the final essay, which was an extensive conversation with an LLM, as well as the Blockchain chapter.)
I appreciated Steyerl's focus on labor and impacts of the technology, as well as the more philosophical explorations. Some of the essays felt a bit disjointed, jumping around unexpectedly -- she's making a lot of cross-connections so some work better than others.
I liked the idea of "data populism" as encoded social bias (in Mean Images). That her prompt of "protest" produced images with the focus on riot police was not surprising, but revealing.
The other day I thought “huh, haven’t heard anything about Hito in a while” and wouldn’t you know it, she just published a new book.
A welcome collection of essays that bring Steyerl back into an immediate relevance (as the essays in Wretched of the Screen and Duty Free Art have understandably become somewhat dated already). If you are looking for some compelling, and at times peculiar, insights into life among AI, this is worth checking out.
Started out strong and was maddening, mind blowing, extremely disturbing, and smartly researched. As artist writing, predictably, I did find Steyerl sometimes too stuck on a particular metaphor or word play - hot/cool, entropy etc - in a way that undercuts her point. (This is a common complaint of mine in any writing, it's something that I find cringey). The essays just petered out around the last three which costs this a star.
Los primeros tres capítulos y el 5 son interesantes. La hipótesis del cambio de paradigma que supone la producción de imágenes con generadores de IA (de lo óptico a lo termodinámico) está bien explicada especialmente en el capítulo 2, "Imágenes quemadas", que considero el mejor del libro.
Really enjoyed the book, very interesting ideas however the last chapter I hoped for more. I found the last chapter to be a bit random and maybe unnecessary in relation to the final statement. Other than that love it!
Stopped reading when I saw in the footnotes that the author – and the editors at Verso, apparently – genuinely believe "AI" is synonymous with linear regression. Embarrassing.