The Democratic Socialists of America currently does not have a fully fledged program on artificial intelligence. What it does have is elements of an immediate program addressing AI, which includes opposition to bailouts of failing AI companies, opposing the construction of AI data centers, and signalling opposition to the AI surveillance state being developed by the federal government and private corporations such as Flock and Palantir.

What DSA is lacking is a complete program in Workers Deserve More that addresses AI in full. DSA does not currently have a consistent line on what elected socialists should advocate for regarding AI, how a democratic socialist government would respond to the AI industry, and how exactly this technology would be used in socialist society for the betterment of all humanity.

It is unfortunately common for DSA members to have only a limited perspective on AI, in its capabilities, in its ongoing usage, and in the dangers of the industry. Many DSA members fail to grasp how advanced the latest models have become; how deeply entwined AI has already become with our economy, and the sheer scale of power AI companies already have and continue to accumulate.

Many DSA members only have experience with AI in opposition to its usage in the arts, or witnessing the adverse damage it has done to education at all levels. It is, of course, for the better that DSA members generally do not outsource their thinking to LLMs. Organizationally, DSA also prohibits A-generated artwork in its messaging, which is definitely worthwhile due to much of the working class having an aesthetic revulsion towards them.

Yet the broader technology of AI is clearly far more than a plagiarism machine, even if this is currently the most common usage of it in the public. AI is capable of performing countless tasks that could be automated, massively reducing the amount of labor needed to sustain society, and helping humans achieve scientific advancements necessary for us to combat climate change, end global inequality, and take us to the stars. It is also capable of creating a dictatorship more repressive and terrifying than anything prior, fueling an endless machine of war and genocide to permanently entrench class divisions and the power of a microscopic caste of billionaires.

The Democratic Socialists of America must build a fully fledged program on AI. It critically must have both short-term proposals that stymies its excesses under capitalism and a long-term framework that excites the working class by utilizing AI to build a classless society. I call for the foundation of a working group to write this program, which we would aim to pass in the 2027 convention and incorporate as a part of the DSA’s core program.

Many DSA candidates and electeds have already put forward proposals that we should mirror in our short-term program on AI. It is immensely popular among the general public to dispel nationalistic fearmongering about “losing the AI race” to China, and call for common sense regulations on AI. Our short-term program on AI ought to fight for the common dignity of workers. They should be afforded transparency from AI firms, so they can be able to refuse AI infrastructure being forced upon them, and for AI not to train on their work and data without their consent. The working group will review which legislation enactable in capitalism can be done and come to agreement on what should be done; whether it is regulatory agencies, nationalizations, the creation of state-supported AI labs, or none of the above.

The long-term program I propose the DSA should have for AI, conversely, would contain elements that have not ever been communicated by any mainstream politician. The ultimate program should not merely be reforms to soften the blow of the AI boom, but for AI as an apparatus to maintain capitalism to be entirely negated.

The governance of AI shall transition completely from private enterprises and from closed-source models to public, democratic governance and free and open sourcing. The conflict and contradictions between labor and the owning class that governs AI shall be erased by the abolition of the bourgeois and the self-management of AI by the proletariat. Universal basic income is oftentimes proposed by liberal AI policymakers as a radical economic policy in response to mass unemployment. As socialists, we ought to go one step further and potentially advocate for a transition to a moneyless society where all basic needs are met without any monetary exchange. AI should be used to help transition the economy to the upper stage of communism.

I expect a host of debates to occur in this group, which have already begun among DSA and the international left. Some proposals, such as forcing AI companies to foot their own bills for their financial failures, I presume will not receive much pushback. More controversial issues include how the DSA should relate to the intellectual fields of AI safety and existential risk; both of which I am strongly skeptical of, but it is a discussion worth having. The exact nature of our elected’s messaging, the minutiae of our class analysis of AI, and our specific bases of recruitment on AI-related organizing I hope will be deliberated through democratic discussion.

Membership of the working group will be open to any members in good standing of the Democratic Socialists of America. Meetings and discussion for the working group will be done on Discord; most likely Mountain Caucus’s server. I do note that this working group is open to all DSA members, regardless of caucus affiliation or lack thereof. You can apply to join the WG here.

A Self-Criticism Of My Prior Efforts

About five months ago I published a piece titled “The Anti-AI Manifesto,” which I have not seen discussed anywhere.

In the past few weeks, I have come to the realization that I have been fundamentally mistaken about what I ought to accomplish with the piece. This is an epiphany that I had to make myself rather than anyone pointing it out to me.

It has little to do with the actual content of the piece, which I still ideologically stand by in almost its entirety, with one area for improvement. What I realize is the fatal flaw with this piece's existence is that I have written it with an inverse logic on what I should have really done.

Before I made this piece arguing that a socialist revolution (violent or nonviolent) is the only way artificial intelligence can be ensured to create a better society, I should have at first concentrated my efforts on ensuring that socialists themselves are ready to explain how AI can address the negative contradictions of glass society. In my own context as a citizen of the United States, I should have at first pushed for the Democratic Socialists of America to adopt a fully fledged program on AI, and to have used this piece as an explanatory tool on an internationalist scale. In essence, I finished this grand piece on the future of the socialist movement before trying to ensure the socialist movement from within is what it should be.

There are a multitude of reasons why my recommendation for those in agreement with the basic thought of this piece to get involved in socialist politics fails without socialist parties having a proper line; with both a short-term program opposed to the AI industry in capitalism and a long-term program on how the technology of AI can be used for the working class.

The first is that the working class naturally would only gravitate towards a proletarian alternative to the panopticon of our entire economy being controlled by a few AI billionaires, if the party is indeed dedicated to accomplishing this alternative as part of its fundamental points of unity. Another is that organized socialists should be both consistent in their messaging and informed in their articulation of how the benefits of AI can be used in socialism. While there are DSA members who clearly are capable of explaining what should be done to AI corporations in the short term, and how a moratorium on data centers is necessary, most DSA members struggle to make it clear to everyday individuals how AI can lead to their standards of living dramatically improving.

This rather basic revelation is one that comes to me embarrassingly late. Perhaps I should have thought about how the Manifesto of the Communist Party was not a singular effort by Marx but a carefully agreed-upon document that articulated the central ideology of the Communist League. I did not originally title it as a manifesto, nor did I set out to create one until I was nearly done with the piece towards the end of 2025. But by the time I completed this piece, I knew that it was one and that no title without the word “manifesto” would suffice.

With this having come to me, it becomes clear that my priority ought to be the adoption of a fully fledged program on AI by the DSA. Perhaps I had avoided attempting this step first due to my views trying to avoid American chauvinism; the text of the broader piece is mostly agnostic on the reader's national origin as it's three recommendations are universally applicable. Having the DSA adopt a national program on AI, however, could result in left-wing movements elsewhere adopting their own programs in the same regard.

I also want to reflect on a part of the work that I likely will alter in the short term future; besides the draft program which I had always intended on revamping even before publishing the piece.

In the manifesto I criticize the pre-existing AI safety movement; on the basis that existential risk arguments distract from risks posed by the AI industry inherent to capitalism, as well as opposition to the reactionary nature of many AI safety advocates. After some conversations with Alexei Gannon and reading more literature in this field, I've moderated my position from unequivocal opposition to pre-existing AI safety frameworks to some sympathy. I still find most fearmongering about Terminator-esque apocalyptic risk from AGI to be ridiculous. While I have no doubt humanity will end if a DOGE intern connects GPT 3.5 to the nuclear launch system, it is obviously both extraordinarily implausible for this to happen and a major distraction from AI’s more concrete risks.

What I am more amenable to specifically is accepting questions of AI safety in regard to massive concentration of power and compute. In the piece I have already written about the US Department of Defense’s collaboration with AI companies for their imperialist wars, and the potential for the government to implement a surveillance state over its denizens using AI. These concerns can be tied into questions of AI safety just as I tie them to a broader narrative regarding capitalism. Resource inequality in AI computing power enables this to happen, through major actors such as Palantir or Flock rather than nefarious bioterrorists that Anthropic harps on about.

I am overall skeptical of narratives regarding agentic risk: the Hugging Face incident was a rather expected outcome of training cybersecurity models to be capable of conducting cyberattacks and putting them in a sandbox where they are encouraged to do so. Considering how excitement for OpenAI achieving AGI has only increased since this incident, the profit motive for allowing such incidents to occur can readily be seen. What I do acknowledge is the real potential for AI agents to act maliciously on a very large scale by capitalist actors, such as an intentional, coordinated cyberattack on a rival state to the US empire, and that these large-scale operations would be enabled to an extraordinary scale by the lack of human personnel required. Such agentic risk is very real, and shouldn't be immediately ignored.

This is not to say that I've come entirely around to a popular front with mainstream AI safety advocates or left-liberal politicians who clamor for co-optable regulations. My criticism of putting too much stock on progressive politicians and LessWrong derivative think tanks in Geese Magazine I still believe in. My argument now is that socialists should not entirely spurn AI safety conversations, but aim to carve its own perspective that is unique and unburdened by liberal ideology. It is my hope that, through the creation of this nationwide DSA program, this perspective will be carved out, and that the manifesto can properly serve as a universally applicable response to the current AI boom and this age of capitalism. Thanks for reading, and I hope that if you have anything to contribute, you will join the working group.