The Left Is Not Missing the AI Debate -- It Is Defining It
Polls from YouGov and Pew Research consistently show that Americans are more worried about artificial intelligence than excited by it. Coalitions opposing data centers are winning fights across the country. State legislatures in New York, Colorado, and California have found majority support for AI regulation bills. These are not the hallmarks of a political movement that has missed the boat.
Brian Merchant, author of Blood in the Machine and one of the sharpest labor-focused technology writers working today, makes exactly this case in a pointed rebuttal to Dan Kagan-Kans, whose piece in the effective altruist newsletter Transformer argued that "the left is missing out on AI." Merchant does not simply disagree. He flips the premise entirely.
Not only is the left not "missing out" on AI, but it would be more accurate to say that it is "currently winning the debate" over AI in American hearts and minds.
The evidence Merchant marshals is substantial. The 2023 Hollywood strikes, which foregrounded AI protections for writers and actors, drew broad public support. California's No Robo Bosses Act. Progressive environmental groups sounding alarms about data center energy consumption. Artists like Molly Crabapple challenging nonconsensual use of creative work in training data. The throughline is clear: left-liberal voices have shaped how most Americans think about AI.
The Stochastic Parrot Problem
At the heart of the "left is missing out" argument sits Emily Bender, the linguist whose co-authored "stochastic parrots" paper reframed large language models as statistical prediction machines rather than nascent intelligences. Merchant argues that AI industry figures find this framing particularly threatening because it challenges their project at a structural level.
If someone believes they're building a powerful super intelligence, I'm sure it feels insulting to have someone call it fancy autocomplete.
Merchant's nuance here is worth noting. He does not claim Bender's framing captures everything about modern AI systems. He sees it as a grounding force -- a way to limit the rhetorical power of industry narratives about artificial sentience. The models can be complex and capable, Merchant acknowledges, while still being programmed systems with material properties rather than emergent minds.
There is a reasonable counterpoint that Merchant somewhat underweights. The "stochastic parrot" label has, in practice, given many people permission to dismiss AI capabilities wholesale. When someone calls a system that can pass medical licensing exams and write functional software "just autocomplete," they risk being as unserious as the boosters who call it proto-consciousness. Merchant gestures at this tension but ultimately sides with the skeptics, which leaves a gap in his otherwise careful analysis.
Who Gets to Define "the Left"?
Merchant identifies a fundamental weakness in the Kagan-Kans piece: it never settles on what "the left" actually means. The critique focuses heavily on Bender, who Merchant notes is not a leftist. The only avowedly left-wing figure Kagan-Kans interviews is Matt Bruenig, who argues the left should use more AI -- hardly evidence of a movement in denial.
Kagan-Kans never really seems certain as to what he wants to describe as "the left", for one thing. Most of the critique is dedicated to the linguist Emily Bender, who is not a leftist, and the only person on the left Kagan-Kans interviews is Matt Bruenig, who argues... the left should use more AI.
Merchant suspects much of the tech world's understanding of "the left" derives from scanning Bluesky, where users call AI "the plagiarism machine." He pushes back on the assumption that casual dismissals represent the full extent of someone's understanding.
Just because someone calls AI a "plagiarism machine" doesn't mean they don't have any further understanding of the technology. One may think it a corny, reductive way to describe AI, or to articulate a rejection of it, but... it just doesn't follow to assume that's the limit of someone's understanding of the topic.
The Policy Record Speaks
Where the essay is strongest is on the concrete policy record. Merchant points to dozens of state-level bills backed by unions and organizations like TechEquity, covering labor impacts, surveillance, identity protection, and child safety. Meanwhile, the Trump administration has moved to impose moratoriums on state-level AI lawmaking -- a position Merchant describes as profoundly unpopular.
He also dismantles the notion that the left lacks intellectual seriousness about AI's future by cataloging an active debate most AI industry commentators apparently missed entirely.
The biggest indicator that Kagan-Kans piece was either not particularly carefully researched or not written in good faith is that it failed to mention a debate that unfolded over the last many months, read by much of the left, between Aaron Benanav, Evgeny Morozov, and Leif Weatherby, addressing this very question.
Benanav's plan for managing fossil fuel drawdown alongside technological innovation. Morozov's case for experimentation. Weatherby's proposal to automate the C-suite rather than the factory floor. Nick Srnicek's book subtitled "The Fight for the Future of AI." Ruha Benjamin, Alondra Nelson, and Amba Kak appointed to policy transition committees. Bernie Sanders proposing a data center moratorium. The intellectual production is there; the Kagan-Kans piece simply did not engage with it.
Resistance as Understanding
Merchant's most provocative move is reframing rejection of AI not as ignorance but as an informed political stance. The left's project, he argues, is necessarily larger than the right's, which is content to deregulate and profit.
"The left" must confront the entire political economy of AI at once, not just consider the core technology, which at this point is nearly impossible to assess apart from its owners and developers.
This is the essay's central insight and its most debatable claim simultaneously. Merchant is right that AI cannot be disentangled from the corporate structures deploying it. But the implication -- that engaging with AI tools on their technical merits is essentially capitulation -- risks ceding the entire field of practical AI governance to industry insiders and centrist technocrats. If the left wins the debate but has no one who understands transformer architectures sitting at the regulatory table, the victory may prove hollow.
Merchant himself seems aware of this tension. He acknowledges the left "could lean more into a program" of specific policy goals and that winning public opinion has not yet translated into winning in practice.
As of right now, it's only winning the debate in the court of public opinion. In practice, AI companies are doing whatever they want, with the blessing of Trumpworld.
The Bigger Swing
The essay closes with a bold suggestion: if AI companies truly believe their technology is revolutionary enough to reshape civilization, the left should take them at their word and demand full public ownership and administration of AI systems. Forget basic income. Forget leaving Sam Altman in charge. If the technology is as powerful as its creators claim, why should the public settle for anything less than collective control?
If AI is truly the revolutionary force they claim, and it stands to remake the world from the ground up, if it promises to eliminate skill difference and advantage, then forget pittances like a basic income. Forget leaving Sam Altman in charge. Why should any reasonable person settle for anything less than full equality?
It is a rhetorically effective judo move -- using the industry's own grand claims against it. Whether it constitutes a viable political program is another question. Public ownership of rapidly evolving, capital-intensive technology platforms has few successful precedents, and the proposal invites the obvious question of whether government-administered AI would be any less prone to surveillance and control than corporate-administered AI.
Bottom Line
Merchant's essay is a forceful corrective to a lazy narrative. The claim that the left is "missing out on AI" collapses under the weight of polling data, legislative records, labor actions, and a rich intellectual debate that the original Kagan-Kans piece simply failed to reckon with. The left is not absent from the AI conversation. It is, in many ways, setting its terms.
The harder question -- which Merchant raises but does not fully resolve -- is whether winning the public opinion battle matters when AI companies operate with the blessing of a friendly administration and near-limitless capital. Skepticism and resistance have shifted the discourse. Whether they can shift the material conditions of AI development remains an open contest.