A Letter to Geese: A Manifesto against AI and Luddism

About a month ago, I finished writing a work titled “The Anti-AI Manifesto.”

I have published it in two versions: a complete version that is some sixteen thousand words long, and a short version that can be read in six to seven minutes of time. I write this letter for the purpose of describing, as succinctly as possible, the necessity of writing this piece, and a request to those who are already convinced of workers’ liberation to read it and spread it as far and as wide as possible.

Why write this work?

Nearly four years after the beginning of the AI boom, neither a genuine movement nor a substantive infrastructure that works against the AI industry exists. The AI industry is the overall project of capitalists to entwine AI with the rest of the economy, from trillion-dollar technology companies to the endless amount of AI startups that aspire to be among these companies. It is impossible to avoid this process: search engines, music apps, and–it feels like–everything else under the sun. 

You undoubtedly already know that this industry is perhaps the most exorbitantly funded project the world has seen. You have likely already seen immeasurable damage it has already done to the arts, academia, and the environment. And you likely have observed that for as much atomized voices of disgust have been raised against this economic bubble, or as many artworks created in rejection of generative AI, or even the publication of a well-meaning encyclical from the papacy– that seemingly nothing has changed, save for, perhaps, the halting of the construction of some data centers.

Artificial intelligence as a technology is undoubtedly not going to suddenly cease to exist, nor can anyone seriously believe it is completely useless for any practical purposes. While projections of imminent, self-improving superintelligent AI are undoubtedly overblown, it is pointless to deny what AI is capable of doing or will soon be capable of doing. That AI will change this world is inevitable. But I argue that it is not inevitable that AI will change the world in the way that every tech billionaire claims it will. I argue that the entire world is wrong.

The focal point of the work is that a fundamental contradiction has arisen from the AI industry. The contradiction is that the underlying technologies of AI, and the modern world, have proven that all classes, all injustices, all scarcities, and all alienation can be abolished. Yet these technologies have not automatically created a utopia. What has been seen is the opposite: these technologies have made the world more alienated than ever before, living is becoming more and more unaffordable, and societal inequality has never been worse than it is now.

I note that the arduous transformation demanded of society to build socialism will not be magically skipped by a sufficiently advanced artificial intelligence; the transition to the next mode of production will still require tremendous amounts of effort and turmoil by the working class. What these technologies enable, when developed and applied in the context of socialism, is the structural possibility of the higher stage of communism where labor itself is abolished.

I wrote this piece to both describe this fundamental contradiction, and to declare that the antithesis of the AI industry must be anti-capitalist in its nature. Many existing organizations exist that try to present an alternative to the current project of AI. They are hopelessly bound to the economic framework of liberalism and tacit acceptance of the world as it exists, and thus espouse meaningless objections based on AI theoretically causing doomsday events, or an AI torturing everyone who does not help bring it to conception. These organizations either seek to create an ideologically broad coalition that is unable to answer the fundamental causes of private property and commodification, or, worse yet, simply try to outcompete firms with more nihilistic worldviews. 

The movement against the AI industry must be one and the same with the movement against capitalism.

There arises the question of why I have given this piece the provocative title of the “anti-AI” movement, and why I believe its text is comprehensive enough to be declared the manifesto of this movement. It is my intention, by naming this piece the manifesto against AI, to define the question of what it means to be anti-AI, in a position that is genuinely capable of standing against the current hegemony of the expansion of current tech enterprise.

The current definitions of what it is to be anti-AI include blanket rejection of the technology on no real material grounds, individualistic boycotts with no revolutionary potential, or the aforementioned thought that it is a generic doomsday threat. The vacuum of any real movement against AI signifies that there is also no clear meaning as to what being anti-AI is. To be anti-AI ought not to evoke simple abstinence from AI products, nor to blindly consider AI an apocalyptic threat. The phrase anti-AI ought to indicate a clear rejection of the fundamental paradigm that the expansion of capitalism by AI is desirable or inevitable.

What should I do?

In the work, I provide a draft program outlining a proposed relationship between the socialist movement and artificial intelligence that entails the construction of a dual power that works for the interests of the working class and for the fight against the AI industry. The program is proof that a comprehensive program against AI, one that ideally would be created by the masses, can one day be made.

In order for such a program to be implemented, I give three general directives that I urge its reader to follow. The directives- specifically tailored to be applicable by the working class of not just the United States but every country- ask for the reader to join and become an active member of a party that is socialist and democratic, to partake in the organization of labor, and to recognize the necessity of the seizure of state power to establish democracy and fight the most brutal manifestations of imperialism and reaction when they arise.

Given that you are reading this in a communist magazine, you almost certainly already do all of this. The request I have for you is to read, at the very least, the short version of the manifesto; and I ask that, if you agree with its premises, to spread it as far and wide as you can. These directives have been made countless times before, but they have not been tied to the promise of what artificial intelligence in socialism could positively achieve, and certainly have not been made with the context of the obvious contradiction that is in front of the world for everyone to see.

I ask you this especially if you are someone of any presence whatsoever, whether online or in the traditional media. If you make socialist content for any social media platform, be it Instagram, or Tiktok, or Youtube, then point to this political stance, summarize it any way you wish, modify or expand where I have neglected to address. If you are a journalist, for any publication, from the bourgeois mass media to student outlets, then write of this, in any framing you desire. Even if you are of no renown whatsoever, if you have creative ability; I ask of you to make derivative works of its arguments; whether it be infographics, or short videos explaining its key points, or posters to plaster across cities around the world.

In fact, I ask you to be as skeptical, or critical, or even scathing of the work as necessary. I have no pretension that either version of this work is flawless; much of its thought has already been subject to evolution in the years since I started drafting it; and I am sure it will continue to do so. I urge you, in anything you create, to freely criticize any part of it, whether it be my dismissal of existing AI alignment, the dual power I propose, or my specific praise of the DSA as a revolutionary organization. Ruthless criticism must come, not just by my ideological enemies, but those who are similarly aligned with democratic socialism but are not fully convinced of these methods. This piece deserves nothing less than a trial by fire.

I would like to place equal amounts of stress on both a realistic projection of what I believe this piece could actually accomplish, as well as overt optimism as to the reaches it could attain. I have no delusions that this piece will cause the massive bubble of speculation surrounding AI to burst and for the massive investment of capital into the scheme to end. I know that this manifesto, being read by a massive amount of people, will not immediately start a revolution. Nor do I believe that anything tangible will occur in the immediate aftermath of it being published. No matter what I could possibly write, one pamphlet alone can not cause parties and unions around the world to explode in membership, for massive strikes to occur in every economy, and for the revolutionary liberation of the oppressed to be achieved.

What optimism I have, conversely, is for the work to be read by everyone. I believe, no matter how impossible it may seem, that getting millions to read the short version–and thousands to read the complete version in full–is possible. The optimism I have is for this to stoke fear into the countable handful of billionaires that must, for once in their lives, feel fear. The optimism I have is for the three-word phrase I coined in it - Join, Strike, Fight - to go down in history. I hope that this piece inspires a hundred other works of superior value to be made, and for these works to contain what words need to be said that I could not have said, and that these works inspire the actions that really need to be done.

And I know that something has to be done. For some time after I finished writing the piece, I tried to get away from it all. I tried to forget about the AI industry and how it has seized the world. Writing it has consumed so much of my life that I did not give myself much time at all to breathe, to remember a world exists outside of this struggle. I tried to find the glimmers of creativity and passion that existed in this world; even if I know capitalism has infected every inch of it, surely there was a time when there was less of it all. Of course, it was unignorable. I saw it no matter what site I went to. I saw it on billboards, talked about by people. I saw it in my own life; I saw it in all other lives, in the burning of the wrong Amazon, in the melting of the wrong ice.

I don't want to have to carry the burden of this piece for much longer. I'm tired of reckoning with the world. But I need to keep fighting. I need this dark age to end. I need the hopelessness in the eyes of everyone I ever loved to be gone. For everyone I ever knew who has fallen into the nihilistic drive for capital. For everyone I have never known, who simply wonder about what life looks like beyond the sea. For all of them to be free.

The link to the manifesto is here. You can read either version first, but the long version is what I had originally written and should answer most lingering questions.

For the Geese

I want to take some time to specifically give thanks to Geese Magazine for being one of the few outlets on the Left to critically appraise the technology of artificial intelligence; not just in criticism of the industry’s expansion and questioning the dominant narrative of how it will reshape the economy, but in optimism of how artificial intelligence can be used in socialism, and how the left-wing should neither blindly embrace technology nor reject it and pretend that it does not exist or can’t be used for the socialist cause. I’ve read Geese Magazine’s pieces on AI for some months now, and I thank its forthrightness for going against the individualistic grain of many of the left-wing who fail to understand that AI has use-cases beyond plagiarizing artwork and cheating on homework.

Some detractors of Geese have accused its writers of opportunistically over-focusing on AI and issues of cultural hegemony as desperately clinging to current trends of this time, but I find this to be an antiquated train of thought. Comparisons and rough analogies can be made, but artificial intelligence’s impact on the global economy will not be the same as any technology prior; whether it be the internet, the mass-production line, or the cotton gin. AI is unique in what it could do for the working class. If communists want the masses to abolish the present state of things, they must answer what the world has consistently presented to them as the future. We live in the twenty-first century, and Geese acknowledges this.

Yet, despite my appreciation for what Geese has written regarding the potential of what AI can mean for socialism, I have a great deal of disagreement with its practical thought on how socialists should respond to the AI industry and how it should combat its major tech companies. Geese greatly suffers from placing too much faith in cooperation with existing infrastructure that is critical of the AI industry; and lacks the conviction necessary to refute its foundations on the fundamental level.

I argue in the full piece that an incredible amount of work is needed to be done by the working class, both in the broader socialist struggle but also the struggle to create new institutions that work entirely outside of existing understandings of how AI is to be governed and its exact role in the economy. Conversely, Geese has previously argued that there exist popular doctrines of thought and organizations that can readily be reoriented to support proletarian causes, and that outreach towards these voices can be fruitful. 

Some writers for Geese believe, for example, in cooperation with the existing AI alignment subcultures that generally coalesce around calls for greater regulation and transparency of research laboratories and tech corporations. To this end, some Geese contributors have specifically sought to reconcile Marxist ideas with modern philosophies in AI politics.

Most notably in the writings of 1000 Means Editor and Geese contributor Alexei Gannon, many aspects of “rationalist” thought have been addressed through socialist lenses, such as the existential risk of artificial intelligence and the complex task of aligning models towards human values and preventing their abuse and misuse. These questions are answered with an olive branch and are accepted as worthwhile to be considered in the first place. The x-risk and doomsday scenarios of advanced AI are posited as addressable through the reforms of the state, and the ethical alignment of AI is raised as a problem that can not simply be answered by moral capitalists but by a transition of power to the commons. 

My opposition to this cooperation is twofold. My first concern is the inherent difficulty in attempting to reconcile with a subculture that has historically been so closely connected with brazenly reactionary techno-libertarianism. Even before the inception of the AI boom, intellectuals fixated on the safety of AI have demonstrated obsessions with practical uses of cryptocurrency, the hyper-capitalist doctrine of effective altruism, and racial differences in IQ. Attempting to solve the AI problem on their terms is a dead end.

The second is that many of the tenets of rationalism are both odious to the working class and unhelpful to their liberation. Only a minority of people may be able to identify Eliezer Yudkowsky by name, but they have heard the alarm he has spent over a decade sounding about the dangers of AI; and it has so far failed to do anything but further reify the narrative of AI companies that they are edging closer and closer to a technological singularity with every billion dollars they burn through. At a grander scale, ceding ground to these individuals shifts focus away from materialist analysis of the industry and towards a technological analysis that is not conducive to revolutionary change.

This is not to say that it is entirely impossible for anyone in this circle of the internet and the broader technology superculture to be swayed to the left. I’m sure that there are self-identified rationalists, academics in the field of alignment, and perhaps even current machine learning engineers of prestigious technology firms that might be swayed to the cause of anti-capitalism. Undoubtedly, these individuals bring a vast amount of technical expertise to the table that would be required for us to build any system of substantive value. Yet, I believe that we must be wary of ideological capitulation to the un-materialist ideals deeply ingrained in that group, and they must drop any vestigial romanticism of classical liberalism before they are to collaborate with us. We are communists and must remain as such, no matter how rare this perspective is in the field we make inroads in; and so shall they become.

I also take issue with much of the optimism for the incremental reforms of left-populist politicians. I do agree with supporting their short-term efforts to halt data center construction and prevent the real possibility that the federal government might try to prevent the incoming burst of the AI bubble by bailing out unprofitable companies, and I can’t imagine an anti-capitalist perspective that argues otherwise.

However, much more critique of how progressives in the US respond to the long-term risks of AI should be made. As admirable as Bernie Sanders is for bringing socialism to the forefront of American consciousness, his meetings with “AI safety experts” resulted in him regurgitating trite apocalyptic narratives, and his purely symbolic proposal for the government to take a 50% ownership in the major AI corporations is riddled with glaring flaws. Some of them include the question of how the government would go about taking this amount of control over the companies; if it would do so by compensating its opulent owners with even more capital; and the mere fact that the federal government is currently not a democratic socialist republic but rather an oligarchic mass that could foreseeably just use this ownership stake to support venture capitalists or other anti-popular agendas.

And this is not even accounting for the full-hearted pushes by 1000 Means and Geese for the left to embrace progressive politicians that address AI. Oftentimes, these are made with what I can only really describe as a projection of what politics they want to see in them. It’s too early for me to make a judgment on the merits of an Alexandria Ocasio-Cortez 2028 presidential campaign, but I do sincerely doubt she intends to use the Defense Production Act to forcibly socialize AI laboratories and transfer power over them to labor unions, as Gannon has suggested to me she should do. I also find myself perplexed by how glowingly Nik M. covered Saikat Chakrabarti; someone who is entirely uninvolved with the DSA and has not even declared himself a socialist, although I do agree his given platform was fascinatingly forward-thinking. For a review of American communism, I do wish some more self-described communists were in review.

Although the exact term “dual power” isn’t replicated in the thought of Geese regarding AI, it is my understanding that Geese similarly believes that the power of AI should be diverted from private interests and to the socialist state and democratic, working-class organizations. What I believe is the failing of Gannon and others in Geese is a misunderstanding of how much of the dual power has already been built, and the sheer scale of what needs to be done for it to be constructed.

It is my contention that dual power needs to be developed independently of what currently exists to address AI–not only because they have already failed to achieve much in particular, but because of what dual power will eventually become as we build the socialist mode of production. Eventually, it will need to graduate from just providing an alternative to current AI companies to totally eclipsing them. If we are going to address class society, then we must be radical and not be dragged down by the vestiges of the previous order. I find it improbable that we will find success by adopting and co-opting structures that, right now, barely dare to imagine a better world. These are simply attempts to try and ensure that the current world doesn’t self-destruct

Lastly, I do want to address a piece by Geese that is critical of AI and its possibilities, “ChatGPT Does not Dream of Labor” by Julia P. This piece seems to misunderstand the abolition of labor. It has already been responded to by Labor & Leisure, which correctly points out that Marx has indeed written about the concept contained in Julia’s criticism of AI, and that subjective attributes such as the “enjoyability” or “soul” of an occupation should not be determinative if they are fit to be automated. I think that this response does go too far in suggesting that opposition to automation is purely out of desire to create a better bureaucracy rather  than one that is eliminated. While that may be the case for some forms of work, there are other occupations being automated right now that have legitimate reasons for them to continue to be performed by human beings. Perhaps those who oppose the abolition of labor might simply misunderstand what work is. The abolition of labor in communism will be the abolition of time dedicated to labor. However, the resulting free time will not merely be idle time, but will also consist of productive time done completely voluntarily and without the need for any form of  compensation.

I hesitate to call this a lack of imagination on the part of Geese’s contributors, as Geese’s vision of a socialist future indicates no dearth of the ability to imagine what could happen. I do not see any inability to defy bourgeois “common sense” projections of how the coming years and decades will appear for socialism in its writings. Rather, I attribute this deficiency in independence to a terminal lack of boldness.

I believe socialists must be bold; not just in their theoretical vision of society but in differentiating themselves from the plans of milquetoast social democrats and capitalist-friendly reformists. In urging the Left to tail progressive Democrats and LessWrong veterans with the hopes that engagement with them could potentially sway them to the cause of democratic socialism, it is my opinion that Geese promotes a counter-productive strategy. This approach weakens our ability to clearly present a revolutionary alternative for the working class rather than simply a slightly more humane remedy to its ailments.

In many ways, I believe this frustrating tailism is emblematic of a wider problem with the thought of Geese and the more popular-frontist wing of the DSA in general. In Geese’s approach to politics, I have perceived a frustrating lack of boldness in hoping the DSA will eventually break from the Democrats as an independent workers’ party–and even a desire to resort to attention-grabbing measures that tie us further to them.

But ultimately having these electoral debates is highly worthwhile; and I see merit to much of Geese's arguments even if I wholeheartedly disagree. I won't mind in the slightest if the readers and writers of Geese continue their popular-frontist outlook if they take a harder line on the AI sphere's moderates, and are more willing to make revolutionary, defiant calls for what should be done to the industry.

I have specifically hoped for my work to be disseminated in Geese Magazine, no matter how frequently it has already discussed AI. It is the culmination of my best attempt at synthesizing the great breadth of thought Geese and countless others have had on what is to be done. Individually, every aspect of it, save for perhaps some novel aspects of the draft program, has been written about elsewhere. But there does not exist a singular place where, in an easily digestible format, what needs to be said about AI is said. What I have requested above–to read this work’s thoughts, engage with what you disagree with, and share what you have agreed with- applies doubly to the readers of this magazine.

I lastly hope that this piece breaks the paradigm where there is only full-hearted embrace of artificial intelligence as it is developed now, or the neo-luddite dismissal of the entire technology as an affront to society. I have only observed a few leftists who have decided that the industry is somehow progressive, perhaps from an accelerationist perspective that AI will spontaneously become anti-capitalist. It is, however, frighteningly common to see individuals who wholeheartedly do not see any future where AI can be used to create a better world. They obsessively focus on futile endeavors, such as personally avoiding the use of generative AI products and shaming those who do, rather than any practical organization.

The lack of any clear or substantive definition of what it means to oppose the trillion-dollar industry ultimately means that the opposition to the hegemonic worldview is ours to be defined. The left should be incredibly optimistic about the potential for what technology could achieve, and relentlessly critical of what it is currently used to do. The left must develop a clear program of what will be done so AI liberates us, rather than developing an AI surveillance state and further destroying the arts and academia. And to do this, the left must point to the clear contradiction of AI and use it to energize the masses into joining us, striking against their slavers, and fighting for a universe that is free.