Your life is larger than an LLM

“There is nothing more mysterious than a TV set left on in an empty room,” Jean Baudrillard once wrote¹. “It is as if another planet is communicating with you. Suddenly the TV reveals itself for what it really is: a video of another world, ultimately addressed to no one at all.” No generative artificial intelligence system has suddenly revealed itself to date, but their outputs ultimately don’t address anybody either. AI can compute pretendedly, but it cannot know who we are. 

The unconsciousness of large language models (LLMs) has not halted their release into realms of conscious life. Increasingly stitched into our social fabric, new tools are causing tension at work and in leisure. Some resist new capabilities, others find them revolutionary. Like the TV left on in an empty room, AI emits glows that few directed and that no one can agree on how to define. 

As an advocate for deeper community involvement in AI development, I recently spoke with Berlin based sociotechnologist Jürgen (tante) Geuter about the importance of collective action. Having studied philosophy and computer science, Jürgen wants us to remember that tech and its advancement is not what matters most, we are. Innovation alone is hollow without our meaning. 

Turn your attention away from the screen 

Kathryn Carter: You describe yourself as an antifascist, feminist and luddite. How do the values that underpin these movements shape your visions, demands and criticisms of current technologies?

Jürgen (tante) Geuter: They’re distinct perspectives, but the value of conviviality connects all three. Feminism helps us look outwards. Antifascism helps with analysis and with looking at what needs to be done; it has a long history of standing against inhumane structures. And Luddism is useful for framing how technology could destroy livelihoods and the social fabric. Connections between politics and technology is what the original luddites symbolise; they fought what they referred to as corrosive technologies. Today, we can see examples of corrosive technology every day, but we rarely connect them to bigger political values. All of these perspectives allow me to apply historical and political concepts to current seemingly merely technological structures.

Kathryn: Speaking of concepts, as Research Director of ART+COM², you develop new forms of expression for the computational future. What kinds of communication models protect human autonomy in the midst of what some scholars now refer to as big tech authoritarianism?

Jürgen: At ART+COM we never build purely digital artefacts. We believe all experiences should be connected to space, to physical locations, and the meaning of those locations. Contrarily, a lot of the digital technologies we use today are about abstraction, alienation; they remove you from everything that matters. They don’t allow us to look at world, instead we look at someone else being in the world through screens.

Kathryn: The French theorist Jean Baudrillard once said³ that the image “bears no relation to any reality whatever: it is its own pure simulacrum.” He believed that such representations of the world are “no longer in the order of appearance at all, but of simulation.” Leading perhaps to the alienation you speak of. 

Jürgen: Yes, and that doesn’t mean that all media is bad, but we need more technologies that can help us live better lives. Today we see a lot of personalisation, positioning people in their own little isolated bubbles. We need to use tech to get people in contact, and not just with the people algorithmically chosen for your feed. We need to figure out how to coexist with all people we share space with. There may be conflicts, but we need tech that facilitates that kind of real life connection, because that’s often when important discussions happen. Meeting people is not about pressing buttons. We need to bring back the joy of building a life together as communities. 

Take your contribution seriously 

Kathryn: Peaceful coexistence is something no large language model (LLM) can understand, yet many still ask LLMs about real world issues. Not you though. On your website you make the statement⁴: “Nothing I write, present or otherwise attribute to myself is generated by or created with the help of so-called “AI” systems. Can you explain why you feel this way about “AI”?

Jürgen: There are many reasons why I critique these systems and why I refuse to use them. My very public refusal is influenced by Luddism. Because what Luddism teaches people is that you can say no. You can just say no, this will not happen. This is me saying that, at least in the parts of the world that I can control. I write to think, to explore my thoughts, and to connect with other people; I don’t write because I need a text artefact to be produced. My writing is not that good, but it’s good enough for me to communicate; to be in conversation. And I don’t want a conversation that is meaningless, I don’t want to talk to a random number generator, that’s meaningless. Writing and reading is talking, in a way, and I want to talk to real people and to their thoughts and experiences. 

Kathryn: Many people now resist AI, but others still entrust a lot to LLMs. What would you say to individuals who feel AI has improved their processes, intellectually and creatively?

Jürgen: I don’t need my thoughts to be filtered through whatever Anthropic thinks is “right”. I want people to write what they care about. And if that’s written in a “bad way” that’s fine. I’d rather read a badly written something full of typos that is true than a polished thing that is meaningless. If it looks like shit, it doesn’t matter, because you did it. That’s my main criticism of these systems, that’s where my emotions run the highest. Yes, you can polish your thoughts with AI, but now they are no longer your thoughts, this is no longer communication. 

Kathryn: Some choose to vocalise their reluctance (and refusal) to use AI only in secret. Why do you think people are so afraid of admitting their preference to not adopt AI in their practice?

Jürgen: It’s still seen as the thing to do, it’s the innovation and you need to show that you’re on board. So people sometimes tell stories of “reasonable use” to claim the middle ground. If you depend on having a job, you need to show you’re with the program, that you’re skilled. And these days showing that you’re skilled does not come from producing good work, [instead] you play the game. You say oh yes, I know AI is the future. And then in two years, it may collapse. 

Kathryn: Like the collapse of the Metaverse? Approximately $73 billion was invested in that endeavour apparently, only for it to be abandoned⁵. 

Jürgen: Yes, only a few years ago people talked about the Metaverse as though it was a reasonable thing; now it’s AI. Western societies in particular have replaced thinking about the future in social terms—in terms of what we need or what we want—with stories of technological progress. 

Kathryn: The late French critic Paul Virilio believed that the embracement of speed and agility resulted in a propaganda of progress⁶. His theory feels particularly poignant today, amid the hype.

Jürgen: Saying you’re adopting  AI means you’re part of the future, of the trajectory. It’s as though the world is saying you’ll be left behind if you step away; the mainstream narrative offers security. If you update your devices every year, you’re part of progress. It doesn’t matter if people are getting poorer, if rents are rising, if the climate is collapsing, you’re still part of the narrative that everything will okay if you have the latest tech. 

Kathryn: Stepping away from that narrative requires you to write your own story. 

Jürgen: Yes, if you step out of that, who are you? That’s very confronting. It can be very alienating for people, because often stepping away leaves them with nothing. Even if you know the tech you’re rejecting is bullshit, and that it goes against your ethical values, not following still puts you outside. Then you need to find your own way, and that’s stressful.

Participate in the future beyond your feed 

Kathryn: In 2017, you published a piece on the data-algorithm dualism⁷. “Most of the algorithmic systems influencing our lives are proprietary and opaque,” you write. In your words, algorithms hide internal structures and mechanics, forcing critique to treat them as black boxes. Do you think that the treatment of datasets as trade secrets⁸ impacts our ability to speak truth to power?

Jürgen: I don’t know if proprietary datasets are the main concern. The bigger problem is the idea of the human in the loop. We’re told that AI tools help us and that we’ll then make “better informed” decisions, but we can’t make informed decisions once we’ve involved generative AI. Making choices between output variants requires understanding the logic that underpins them. What led the AI system to make the suggestions it did in the first place?

Kathryn: As humans we can explain ourselves when asked, but AI systems can’t?

Jürgen: Yes, the stochastic systems that we call “AI” today cut that tie. Tools give you suggestions, but can’t tell you why. This has nothing to do with the proprietary nature of the data. As a computer scientist, I can tell you that the promise of computers was that they are fast and that they can make fewer mistakes with certain tasks. This is all useful, of course. But the idea was always that you had an algorithm you could review. And if the computer gave you a false response, you could dive in and understand what was going on to fix it. That’s gone now, and that shift away from operational transparency is far more important than where the data comes from. 

Kathryn: So when people talk about deep reasoning models, there’s still no way of knowing the reasons why AI has generated its outputs?

Jürgen: There is no way to understand why decisions or suggestions are made by AI. Whereas people can share the dumbest ideas and still argue why they made the suggestion, allowing for open debate. Tech companies imply that AI tools hold all the world’s knowledge, but even that’s not true, tools are not based on knowledge but on data shaped through mostly a western perspective and aggregated by the western viewpoint into datasets. That doesn’t represent the whole world. You’re led to believe AI tools are “smarter” than you are, but they’re not. 

Kathryn: Speaking of perception, German philosopher Walter Benjamin once wrote: “Humanity’s self-alienation has reached such a degree that it can experience its own destruction as an aesthetic pleasure of the first order.” He seemed to believe that an abstract understanding of the self as 'an object of contemplation' had supplanted an embodied self of subjective experience⁹. Do you believe the invasive integration of tech into routine existence causes some to act more like avatars than people, causing them to look upon the injustices of the world from a position of spectator, not participant?  

Jürgen: I think the invasive integration is part of it, but it’s the abstraction element too; all computers and maybe all media systems do this. We get removed from everything via the design of interfaces, and everything has the same quality. Whether it’s about war, sport or cooking, it’s all just content we’re churning through, and it weirdly starts to lose its shape. Types of media once distinct—radio, film, television—are now melted down into “feed stuff” that people slurp up. So yes, this leads to people behaving like spectators, because consumption of content is the only mode of action you’re taught to take. The landscape facilitates a “lean back” mode of existence. 

Kathryn: The Atlantic recently reported¹⁰ that people with ties to Silicon Valley are now “insisting that the AI race slow down and the technology be overseen not by executives and engineers, but by the broadest population possible.” What practical steps can people take if they want lean more forward, to participate in shaping how technology is designed and deployed in future?

Jürgen: It depends on the tech. With AI, it’s interesting to see how unpopular it is, and to study how that happened, because I haven’t seen this kind of response in my lifetime. So opening up the public discourse, so that people can properly express their concerns, that’s important. And that’s also why I think it’s so important to state: no, I will not participate in this, because it’s wrong. It feels very small because we’re not in positions of power, but it still matters. We do empower people when we point out, as the luddites did, how technologies such as AI can be corrosive to the social fabric. It doesn’t hurt to say: I don’t use this. And find other people you can build a movement with; there is value in that. 

Consider the process not the product 

Kathryn: Russian revolutionary Leon Trotsky once wrote¹¹: ”Art, is it said, is not a mirror, but a hammer: it does not reflect, it shapes. But at present even the handling of a hammer is taught with the help of a mirror, a sensitive film that records all the movements. Photography and motion-picture photography, owing to their passive accuracy of depiction, are becoming important educational instruments in the field of labor.” Today’s equivalent medium of concern is generative AI. How do you think its existence impacts the movement of hammers of art?

Jürgen: You’re no longer really creating art, most of the time, when you engage AI. What we see are people prompting systems that then generate deliverables that looks very “done”. Then the individuals post-rationalise that the output is what they wanted. But they don’t go through the process of experimenting, changing things, figuring things out. So, in a way, generative AI systems claim that you don’t need to go through your process. 

Kathryn: That’s quite sad, it feels like a loss of sorts. "In many ways writing is the act of saying I," Joan Didion once said¹². If the process is thieved by tech, what happens to the psyche?

Jürgen: Everyone has a different process, or maybe even processes, depending on your mood, or the time of day, or whatever. We contain multitudes. But AI systems claim: no you don’t, no you don’t need to have a process. It’s all about the output, and that’s related to the metrics-driven capitalist logic of not needing to create for humanity or expression, but so that you can have something to monetise. 

Kathryn: Outsourcing the intellectual or creative process is sometimes described as a productivity boost, due to time saved. But what’s the sacrifice? 

Jürgen: AI tools promise: you can have something without having to put in the work. That’s tempting, of course, because work can be hard. You’re a writer, but writing is not pleasant for me, I just know I need to do it. I like having written, not because it creates an article, but because I’ve experienced myself, my thoughts, my views on the world, in the process. For those of us who create to understand our world more deeply, outsourcing the process takes that away from us. 

Kathryn: You sometimes hear AI enthusiasts lamenting the blank page, and that sensation of staring at it without knowing where to begin. But to penetrate that blank space with words grown from your imagination, that’s quite special, even when it feels difficult. Yes, AI can draft entire documents in seconds. But I still struggle to understand the value of removing ourselves from the process by passing it over to a computer program. Time is saved, but soul is lost. 

Jürgen: The German word for understanding is begreifen. And begreifen means to grip something, it’s the literal meaning of the word. It means hold it in your hand. That is very meaningful, to me. I need to hold things in my hand. 

Kathryn: I love that, it implies that there’s a tactility to knowing and wisdom, which is quite beautiful. When you foster begreifen, you embody something. Knowledge is not just information stored in the mind; knowledge is held by the body. 

Jürgen: Exactly, but technologists, and of course corporations, have a bias towards doing things more quickly. Less hours means less time they have to pay people for. It’s cutting jobs. 

Kathryn: The sad reality of capitalism. 

Jürgen: Technology and capitalism fit together quite well. Not all the time, but often technologists are dragged down the path of productivity and efficiency as priority because they never learn to challenge those ideas. When you study computer science, you learn about abstraction. You learn to build systems that connect to real world processes, so you become skilled at abstracting what is tangible to computerise scenarios. What I’ve found is that many people in the field start to believe that because they can abstract anything, they can understand anything. The attitude that follows is: the way I understand the world is the way the world needs to work. And that’s not ethical. 

Kathryn: Joseph Campbell once said¹³: “Now I am rather an authority of gods, so I identified the machine—it seems to me to be an Old Testament god with a lot of rules and no mercy.” What are the risks when computer scientists, and their rule based mindsets, start to influence the societal landscapes shared by humans who, as you say, contain multitudes?

Jürgen: When we force everyone into preprogrammed boxes, preprogrammed structures, treating them like machines, we can strive for efficiency, but the approach deskills people who have ideas. Many computer scientists lose sight of what they’re supposed to be supporting—people, not just productivity. So, in the tech world, you lose connection to what matters sometimes, yet you still feel like you understand everything. The worst mansplainers in the world are computer scientists, I am deeply sure of that. 

Kathryn: So when the handling of the hammer of art is taught with the help of generative AI? 

Jürgen: To return to, and adjust, Trotsky’s hammer analogy differently. Computer scientists have the hammer of I build systems. So every problem becomes something a hammer can solve. 

Remember your moral compass 

Kathryn: You’re someone who believes that technological innovation can help create a better world, but only as long as the inherent tendencies towards exploitation and discrimination are being addressed. How can we transform the trajectory of tech to empower people in future? 

Jürgen: There is a lot of work to do, and a lot of it is just figuring out the problems. Once we grasp the problems fully, we can start to make decisions and present solutions. In the meantime, as I’ve mentioned a few times, we can teach people to say no, that’s more important than people realise. Saying no gives people agency, it allows people to understand they can make up their own minds. As soon as you take the ability to participate—to say yes or no—off the table, you’ve lost agency. 

Kathryn: The Russian author Aleksandr Solzhenitsyn once said¹⁴ that “a stable society is achieved not be balancing opposing forces but by conscious self-limitation: by the principle that we are always duty-bound to defer to the sense of moral justice.” His words, like yours, remind us that every little decision we make for ourselves, by our own moral compasses, contributes something. 

Jürgen: With conscious attention, we need to reshape our expectations and view of technology. Currently, the future is not being planned in accordance to our needs, but to serve technological progress. We hear words like newer, faster, better, but it’s all about tech, not people. We need to drag people back to thinking about technologies as political entities; technologies as things we need to align with our values. When we return to conversations about what we actually want for ourselves, we stop ourselves from being dragged around by narratives of: this is what technology can do, so therefore we must do it. We don’t have to do anything that’s not aligned with our values, that’s what we need to remember. 

Kathryn: Technologies are often framed as panaceas. One obvious example is AI revolutionising healthcare¹⁵. Mainstream narratives zoom in on the tech transforming treatments, but often omit considerations of what might need to change to decrease the likelihood of disease formation. Housing, for example, and access to essentials for all, may just as powerfully improve health outcomes by preventing illness to begin with¹⁶. What if we took the millions spent training a new large language model and invested it in ensuring everyone had somewhere safe to live instead?

Jürgen: That’s really important, exactly the kind of question that you just outlined. We have plenty of people who present technology as the solution. But we need to take a step back and ask: how many resources do we want to spend on this technology? And what will it cost to actually fix the problem? A lot of tech is released by companies who admit that we still need to experiment to figure out how the technologies will work; the creators of the capabilities admit they don’t know. Yet we do know how to house people, it’s a known solution that works. So why aren’t we doing that? I want someone to explain it to me like I’m a five-year-old.

Kathryn: It seems obvious to many of us that so much of what we need to address today can’t be fixed by LLMs. That said, untainted by the influence of big tech authoritarianism¹⁷, technologies can empower individuals and collectives when developed by human-led design. 

Jürgen: Yes, I do actually believe that. I am not against technology, but what do we really need? And have we asked the people who are in need what they want? We should start by identifying the problem, but right now everyone just wants to use the shiny new tools. Technologies don’t really matter, they are completely uninteresting, and I say that as a technologist. 

Kathryn: “Information can tell us everything. It has all the answers,” Jean Baudrillard writes¹⁸. “But they are answers to questions we have not asked, and which doubtless don't even arise.” Right now it feels like the same can be said for tech that promises so much and yet delivers so little to the people. We must pivot from soulless solutions back towards the human spirit. 

Jürgen: Yes. Right now, we’re making people’s lives harder than they already are with generative AI, largely because of laziness. We should put in work to make people’s lives easier. That should be our job as people who understand technologies. 

About the interviewee

Jürgen (tante) Geuter is a sociotechnologist, writer and speaker who advocates for moving deliberately and fixing stuff. Versed in philosophy and computer science, he is known for his critiques of technological hype. You can explore his work at smashing frames.