Why We Shouldn't Panic About AI - City Journal

TEXT ANALYSIS: "Why We Shouldn't Panic About AI" – City Journal

THE DISSECTION

This is a nostalgic sedative dressed in historical costume. The author recycles the standard "Luddites were wrong, technology always creates more jobs than it destroys" narrative, layering in select vignettes from the steam engine era through the 1960s robotics panic to argue that AI will follow the same benign pattern. The piece is constructed to comfort a specific readership—educated professionals who have been nervously monitoring AI news and need reassurance that their skills remain economically relevant.

The rhetorical architecture is predictable: acknowledge the concern superficially, then demolish it with analogical history, then sprinkle in a few weak exceptions ("AI can't replace plumbers"), and finally reassure that "adaptation will take time." It's a policy-free zone—no proposals, no predictions, just vibes dressed as historical analysis.

THE CORE FALLACY

Category: Category Error on Steroids – Cognitive Work ≠ Physical Labor

The entire historical analogy collapses on a single structural difference the author either cannot see or will not name:

Previous automation waves automated physical labor. AI automates cognitive labor.

The spinning jenny replaced the hand of the textile worker. Factory robots replaced the arms and spine of the assembly worker. These displacements freed humans to do more cognitive work—managing, coordinating, designing, selling, innovating. The post-WWII economic order was built on this bargain: if you have a brain, you have economic value.

AI does not compete with your arms. It competes with your brain.

The steam engine made cloth cheaper, which expanded demand, which required more clerks, managers, marketers, and salespeople to handle the increased commerce. This mechanism presupposed that human cognitive labor remained necessary to scaled operations. When AI makes cognitive production cheap, that demand-side multiplier loses its engine. You cannot expand demand for cognitive workers when AI saturates the supply at near-zero marginal cost.

The author is citing the respiratory system's response to a prior injury to argue that the patient will survive a heart attack.

HIDDEN ASSUMPTIONS

1. That new cognitive work will emerge to absorb displaced cognitive workers. The author admits he cannot predict the new opportunities ("futile at this early stage"). So the entire reassurance rests on faith that something will appear. That is not analysis. That is hope.

2. That productivity gains automatically distribute to displaced workers. The historical examples show employment growing—but whose employment, at what wages, under what conditions? The author romanticizes the 1960s manufacturing "golden age" while acknowledging it employed fewer people than pre-robotics. The expanded employment post-robots was not the same work at the same wages. The "golden age" was itself a product of union power, New Deal institutions, and Cold War politics—none of which the author acknowledges as contingent, or currently present.

3. That human judgment remains indispensable in law, medicine, and management. The author claims AI "only recounts existing knowledge" and cannot determine what should be built. This is a 2022-tier understanding of frontier AI capabilities. It is also the precise claim that AI developers themselves, including the ones the author cites (Altman, Amodei), explicitly reject.

4. That the post-WWII employment-income-consumption circuit remains operative. This is the most critical smuggled assumption. The entire historical pattern assumes that displaced workers can move into new roles that pay wages sufficient to sustain consumption. The DT thesis states this circuit is severed when AI achieves cost-performance superiority across cognitive domains simultaneously. The author never engages this claim. He simply assumes continuity.

SOCIAL FUNCTION

Primary Classification: Elite Self-Exoneration + Prestige Signaling

This article performs two functions simultaneously:

For readers: It tells educated professionals—the City Journal demographic—that their cognitive skills remain valuable and their anxieties are overblown. This is ideological reassurance theater for the class most threatened by cognitive AI. "You are not the textile worker. You are the factory owner."

For the publication's ideological role: It provides intellectual cover for political inaction. If the problem is overstated, no urgent policy response is needed. The article does the work of a lobbying effort without appearing to be one. The "let markets handle it, history proves technology is good" frame conveniently neutralizes demands for structural adaptation—whether in education, taxation, ownership, or social provision.

Secondary Function: Lullaby for the professional class. The author even helpfully suggests that plumbers, therapists, and "personal life coaches" are safe. A curious list—occupations that are lower-status, less economically leveraged, and more resistant to AI specifically because they rely on physical presence and emotional labor. The subliminal message: "The AI won't take your job, but it might take the jobs of those other people. You'll be fine."

THE VERDICT

This article is a symptom of the disease it claims to diagnose.

The author has correctly identified that past technological transitions were net positive for employment. He has fatally failed to recognize that the mechanism of that positivity—human cognitive labor as the scarce factor driving expanded demand—is precisely what AI destroys. He is reassuring readers that the lifeboat will hold water because previous boats held water, without noticing that this boat has no floor.

The historical comparisons are not evidence. They are selected memories that confirm a preferred conclusion. The 1960s manufacturing expansion he romanticizes was achieved with robots, yes—and also with unions that no longer exist, a regulatory environment that has been gutted, a Cold War imperative for high-wage industrial employment that has evaporated, and a labor force that has not yet been tested by AI that can perform at the level of experienced professionals.

The most damaging sentence in the article is the penultimate one: "AI adoption may move faster than that, but its effects—good, bad, and indifferent—should unfold at a pace that will allow people and practices to adjust."

This is the assumption the entire piece rests on, stated plainly as though it were a neutral observation rather than a contestable bet. The pace of AI development is not a gentle wave. The pace of adaptation in legal systems, educational institutions, and political structures is not accelerating to meet it. The assumption that "things will be different this time" is dismissed as unlikely, but the structural reasons why this time is different—the nature of the technology, the scale of cognitive displacement, the absence of countervailing institutions—are never addressed.

Classification: Partial Truth Sold as Complete Truth. Historically Informed Nostalgia Dressed as Economic Analysis. The Intellectual Equivalent of Telling Someone Their Terminal Diagnosis Is Probably Nothing.

VIABILITY ASSESSMENT OF THE ARTICLE'S CENTRAL CLAIM

This article will age like a 1950s reassurance piece about nuclear power having no downsides. The structural analysis is absent. The historical analogies are structural mismatches. The policy conclusions are none. The social function is to calm the professional class and let the people who should be building adaptive infrastructure off the hook.

The verdict: Copium sold at a premium to people who can afford to believe it.