‘AI’ is a tool that, by its developers’ own admission, is meant to enable control of population that subverts, and overpowers, all existing political apparatuses, and which, by design, is meant to substitute people’s capacity to make their judgements. ‘AI’ users and promoters who think they are capable of conveniently plugging themselves into this massive thing and using it to their advantage, without being dominated by it, have arguably more delulu optimism than us who believe that ‘AI’ can (and must) be abolished. If we muster just a fraction of the faith that ‘AI’ boosters allocate to ‘AI’, and put this faith into a revolutionary project instead, that would be enough to change the world. We just need to convincingly argue for such a (re)allocation.
In short, working towards ‘AI’ abolition is working towards liberation. But what exactly do we mean by that?
Abolition means working towards a liberatory alternative.
Abolition as a political project is most famously associated with the abolition of slavery. The political project of abolition identifies an oppressive institution that it wants to eradicate (in this example, slavery); and importantly, also, the project of abolition suggests a positive worldview (in this example, a vision of the future where humans treat each other equally and do not dehumanize and enslave each other). Taking an abolitionist perspective is not (or at least not primarily) being against an oppressive institution; taking an abolitionist perspective is being in favour of a liberatory alternative to an oppressive institution.
Other abolitionist projects follow this logic, too: prison abolition works towards restorative justice as an alternative to the carceral system; police abolition works towards nonviolent public safety as an alternative to the police (and/or policing) state; family abolition words towards community care as an alternative to normatively imposed heteropatriarchal nuclear family structures.
As AI abolitionists, we place ourselves in the context of other liberatory struggles against oppressive institutions and in favor of humane alternatives. AI abolition works towards a revolution from the left as an alternative to global technofascism.
Many proponents of the abolitionist movements mentioned above – abolition of slavery, prisons, police, and family – come to the conclusion that their abolitionist project is not possible within a capitalist system. The same is true for AI abolition. We believe that the abolition of AI is only possible if it goes hand in hand with the abolition of capitalism.
At this point in the text, since, as the famous saying goes, “it is easier to imagine the end of the world than the end of capitalism” (Frederic Jameson), it is worth reminding ourselves that not only are alternatives imaginable; they are also feasible and realistic, as for the most of human history — for the absolutely longest time of our human existence on earth — we humans lived without slavery; lived without prisons; lived without police; lived in communities without cisheteropatriarchal norms; and lived without AI.
While we argue about what “AI” really means, real harm is being done.
Whole books and articles have been written, and will be written, about “what exactly” AI is. While these discussions are interesting, they are also beyond the point for our abolitionist purpose here; arguing about competing definitions of AI is distracting our eyes from the actual prize: an urgently needed revolution from the left. The good news is that we do not need to go through all of the small print in order to start working towards that positive worldview. This is also reflective of reality: since “AI” means different things to different people, you are invited to join the struggle — no matter if you are fighting against…
- AI as a set of technologies;
- AI as marketing term rooted in eugenics;
- AI as a product with for-profit promises that are failing to materialize;
- AI as a misconception of what computers can do;
- AI as a cultural practice of outsourcing our thinking;
- AI as floating signifier;
- AI as a political argument for the extraction and appropriation of vast material resources;
- and the list continues.
What all of these definitions of AI have in common is that enormous harm is being done in the name of “AI”.
[here could be inserted the typical laundry list of: the planet, the data privacy, the enshittification, the resources, the acceleration of work, the inhumane working conditions for data labellers, the damage to education, the detrimental psychological effects on people, the gigantic natural resources consumed, the gigantic dual use notjustpotentialbutactualuse see warfare in Gaza and elsewhere, the makingpeopledependentonithencelessawareof the issues, etc.]
Of course, technology is never neutral. And we do not buy into singularity narratives into “AI” bringing human beings closer to God, or the Devil, or both. So, what makes AI not just “yet another technology, and why don’t you go abolish the telephone and the printer if you hate technology so much?”, is the enormous scale at which the term is being used to do harm in the name of some diffuse technooptimist future. The harm done in the name of AI is much, much, much larger than any of the supposed benefits brought about by AI. And do not fool yourself at this point (just like we do not want to fool you): it is not enough to try to figure out how to “mitigate the harms while exploiting the benefits”. We can see that all of the harms done in the name of AI are integral part of AI; and are integral parts of capitalism. As such, there is no ethical AI, and never will be; and there is no AI abolition under capitalism.
You can take action, now.
It is easy to feel overwhelmed by the magnitude of the issues. Feeling overwhelmed, however, is not the same as being paralyzed. Even when you feel overwhelmed, you still have the agency to act. We collected here a list of actions you can take, from the smallest to the most involved, sorted more or less by magnitude of efforts needed.
Tips to improve your wellbeing
You can abolish AI in your own life. Starting today.
- Most importantly: find your people. And support each other. None of this can be carried by one single individual; we all need community to support and sustain us.
- Stay with the trouble: In a dehumanizing world, embrace the human experience.
- Ask yourself, not “AI”.
- Ask other people, not “AI”.
- Find out more about the topic: learn & disseminate knowledge about the narratives that make AI seem “inevitable” or “unavoidable” even as these same narratives serve to perpetuate oppression, slavery, de-skilling, authoritarianism and ecological breakdown on a planetary scale.
- Find analogue activities to replace your volunteering your time feeding the BigDumbRobot (aka, do more of what you actually enjoy, instead of scrolling through Instagram or conversing with ChatGPT)
- Re-skill and up-skill yourself by phasing out tech crutches (like on-the-road access to your emails, 24/7 online availability for communication, GPS navigation) by working on your planning, organizing, navigation and community-building skills
- Find like-minded people, gather in real life, and organize together
- Stand up and speak up: instead of being part of the discursive normalization, be part of the critical voices who ask foundational questions
- Stretching exercise: “reach up and out”. Reach up to the people you can reach that are the highest up in the decision making hierarchy at your workplace, in your community. And reach out to the people that are not (yet) part of your social environment: talk to acquaintances, strangers, the media.
- Organize demonstrations and strikes (see Pull the Plug in the UK)
- Organize conferences (see Cables of Resistance in Germany)
- Start building resources for a difficult future (community and survival skills)
- Direct action: start blocking and dismantling instruments of oppression
“AI” is human cultural practice.
When we ask to abolish AI, we are not asking to abolish numbers. The models that be may peacefully continue to exist in hypergeometric space, as they have for time eternal, independently of our human dis/usage of them.
When we ask to abolish AI, we ask to abolish what has become (from a 2026 standpoint), over the last few years, a prevalent cultural practice whose most salient feat is its overwhelming uncritical adoption (which, as we explain below, is not a fluke, but willfully engineered into AI systems).
“AI” is the cultural practice of outsourcing our thinking.
What people that praise “AI” for its possibilities “for the benefit of humanity” (solving the climate crisis, eradicating racism, curing cancer, and render the working class superfluous through automation) miss out on is a simple reality check of scales and proportions. All of the enticing future promises so far have largely failed to materialize — quite on the contrary, “AI” is often driving the opposite of what the salvation promise is claiming, i.e., fueling the climate crisis, exacerbating racialized violence, facilitating the mental and physical suffering and death of billions, and trapping enormous segments of the population in increasingly meaningless and agencyless work. And at the same time (here the scale comes in!), if we weigh every single human’s participation in the BigDumbRobotInc. equally, the overwhelming majority of us humans nowadays is using “AI” for one single purpose: to outsource our thinking.
“AI” brings fake salvation promises to the selected few… and massive harm to everyone.
For every single researcher in a white coat trying to “use AI to cure cancer” (and again, let us be real, often an underpaid and underslept fellow whose institution does not grant them enough financial support to hire a student assistant with programming skills, which is why they are prompting a chatbot instead of writing the R script themselves), there are two million insecure and lonely fellow humans asking the same chatbot: “Why are they not returning my texts?”, and: “Is this weird feeling in my stomach sadness or anger?”.
For every single office worker who smugly “uses AI to do their work for them” (most likely, creating a sequence of words that resemble a meeting notes summary, which no-one will ever read, let alone factcheck, from a meeting that could have been an email anyways; but at this point, let us say goodbye to the reading and writing skills of the smug office worker!) there are five million people working under conditions of de-facto slavery (i.e., slaves) in the BigDumbRobotInc.’s supply chain, doing the very real, very material, very much still non-automated and physically and mentally violent work of mining the coltane, assembling the computers, constructing the data centers, recycing the cables, annotating the data sets, filtering the hate content.
“AI” is addictive.
The subtle way that “AI” creeps into our lives to settle, nest, and infest our psyche is best illustrated precisely by this one moment of lulling comfort: “why go through the pain of thinking for myself when i could simply ask a chatbot?” The comfort of having an instant answer provided to any question; and an answer that has been granted authority by society, at that — this comfort is as fake as it is addictive.
Just a couple of years ago, and to be more exhaustive, for the last 100.000 years, we were perfectly capable of life without “AI”. Today, an increasing number of people claim they couldn’t imagine their life without it anymore (as is often the case for technology adoption, see the smartphone, the internet, or GPS). But if we do the mental exercise of zooming out beyond our own lifetime, it is easy to see that “AI” is not a necessary tool for any of our livelihoods — simply because (1) it has never been, and (2) the world has gotten arguably worse since the tool was introduced.
“AI” enshittifies everything.
AI makes things worse: the internet, the news, the tasks at work through fake automation, the education system, the healthcare system, human relationships, the environment, fully automated weapons, art, etc., etc., etc.
“AI” is evitable.
Contrary to current popular techno-optimist and techno-fetishist belief, the adoption of any technological innovation is not willful providence descended upon us from the Heavens; it is our conscious choice. It is our choice whether to adopt a technology; it is our choice how to adopt the technology. And so, being against “AI” as a cultural practice is not the same as being against technology, maths, or the internet.
But why, then, are we being so radical in our choice of words, straight up asking for AI abolition? Why are we so meanly discriminating against the BigDumbRobot? If technology is a choice, and mathematical models underlying AI are not inherently bad, why don’t we then ask for a safe adoption of AI? Why don’t we ask for a safer adoption of AI? Why don’t we ask for critical, transparent, people-centered AI? Why don’t we ask for revolutionary, empowering, ethical, humane, equitable, diverse, fair, just, morally elevating, antiracist, feminist, woke, beautiful, happymaking and allwarsoftheworldsolving AI?
Ok, but first capitalism.
The answer is simple. In the same vein as the popular T-shirt inscription “Ok, but first coffee” — the flirtatious subtext being “I am down for anything really, but my main personaly trait is that I am a coffeine addict” — this is what we have to reply to the proponents of “using AI for the benefit of humanity by building just, equitable, distributed, community-driven distributed AI systems”:
Ok, but first capitalism.
Because let us be real — in the current geopolitical and economic world order, we as a global society are failing at infinitely much easier tasks. And we have been failing for centuries.
No country in the world is anywhere near meeting climate targets (although we have abundant scientific evidence that continuing the status quo will lead to large scale collapse of civilization). Millions are dying of hunger (although there is enough food for all). Millions are dying of genocide (although there are enough courts denouncing the genocide). Billions are suffering gender-based violence and discrimination (although there are enough books convincingly arguing that gender-based violence and discrimination are bad).
One might argue that we are not trying properly. One might even argue that we engineered for ourselves a logic of existence that contains these crises as a necessary evil. And if we cannot solve hunger, war, hate and climate collapse, what makes us think that we can solve AI governace… in a world of hunger, war, hate, and climate collapse? How arrogant do we have to be to think that we who fail at all of the above challenges can use “AI”, which directly constitutes and exacerbates humanity’s problems, in such a smartass way that it will become part of the solution, all while remaining part of the problem?
The only world in which AI could be appropriately used, deployed, legislated, governed is a world free from exploitative power relations. That is, in a world without capitalism. Therefore, we say: Ok, but first capitalism. Ok to all the pipe dreams of “AI-infused” futures; keep up your hopes. But to implement your dream of a beautiful future, do wake up to the fact that such a future is only possible in a post-capitalist word. And first, solve capitalism. Then, we can solve “AI”.
…Sounds complicated, not to say currently unthinkable?
Exactly.
Solving capitalism, while necessary, is likely to take quite a while. And in the meantime, while we’re working on it,
We need to abolish AI. Now.
Because under the current conditions of capitalism the only safe step is to completely eradicate this cultural practice.
Because the more you let a cultural practice encroach on your daily life, the more difficult it is to disentagle yourself from it again — so every day counts (and it is not too late!)
Because contrary to what techno-fetishist logic wants you to believe, progress is not, and does not have to be, linear. The progress that we want to see is the world becoming more, not less, humane.
Because AI is already destroying and extracting everything around us: our climate, our water, our energy, our soil and our capacity to live meaningful lives.
You can abolish AI in your own life. Starting today.
As a contemporary individual, you might feel disempowered to act “against” anything as big as the BigDumbRobot (as cultural practice). Speaking of AI abolition can feel liberating — but also daunting. You can, however, start today to abolish AI in your own life, and then let the waves ripple outwards towards the lives of other people. Here is an incomplete list of feasible steps to take, sorted more or less by magnitude of efforts needed.
- Stay with the trouble. In a dehumanizing world, embrace the human experience.
- ask yourself, not “AI”.
- ask people, not “AI”.
- read more about the topic (insert some resources here); learn & disseminate knowledge
- find analogue activities to replace your volunteering your time feeding the BigDumbRobot (aka, do more of what you actually enjoy, instead of scrolling through instagram)
- re-skill and upskill yourself by phasing out tech crutches (like on-the-road access to your emails, 24/7 online availability for communication, GPS navigation) by working on your planning, organizing, navigation etc. skills
- find like-minded people, gather in real life, and organize together
- stand up and speak up: instead of being part of the discursive normalization, be part of the critical voices who ask foundational questions
- stretching exercise: reach up and out. Reach up to the people you can reach that are the highest up in the decision making hiearchy at your workplace, in your community. And reach out to the people that are not (yet) part of your social environment: talk to acquaintances, strangers, the media.
- organize demonstrations (see Pull the Plug in the UK)
- organize conferences (see Cables of Resistance in Germany)
- start building resources for a difficult future (community and survival skills)
- direct action
- sabotage
- strike
Refuse to accept that comfort is the ultimate goal.
Accept discomfort. Lean into discomfort.
Gain back your real life.
Abolish AI.