TEXT ANALYSIS: Technology Usually Creates Jobs for Young, Skilled Workers. Will AI Do the Same?

THE DISSECTION

This article is an operational prestige piece. It repackages a legitimate MIT study on historical job creation patterns as though it provides useful framing for the AI transition. It does not. The article presents David Autor's research on postwar employment dynamics as though it constitutes meaningful guidance on the AI question—and Autor himself is partially complicit in this sleight of hand by participating in the framing without adequately flagging its irrelevance.

The piece's structural function is to defer the hard question. It performs the ritual of "engaging seriously with AI displacement" by citing expert research, then pivots to optimistic hedging disguised as scholarly caution. "We don't know what it will look like, and who will be able to do it." This is not a conclusion. This is an escape hatch dressed as rigor.

THE CORE FALLACY

The article assumes continuity between historical automation and AI-driven automation. Every prior wave of technological displacement—from farm labor to typing pools to assembly line work—involved machines replacing physical or mechanical cognitive tasks while leaving intact the cognitive infrastructure that humans provided. Routinized cognitive work survived. Bureaucratic judgment survived. Interpersonal context-detection survived. Domain expertise in applying general principles survived.

AI does not automate the tasks humans do. AI automates the cognitive function itself—the processing, judgment, synthesis, and generation that constituted the value premium of new work.

Autor's own research is the autopsy report on his own thesis. He documents that new work "itself gets automated. New work gets old." He notes that the scarcity value of expertise erodes as it becomes common knowledge. This is the exact mechanism by which AI will destroy the new work it creates—faster than any prior technology, and with no remaining cognitive basement to retreat into.

The article treats this as a background observation. It is the central fact.

HIDDEN ASSUMPTIONS

1. Scarcity of expertise is recoverable. The piece treats expertise scarcity as a durable feature of new work, when Autor himself notes it erodes. AI doesn't wait for erosion—it is the mechanism of erosion. Expertise that required a college degree in 2025 may be commodified by 2027.

2. Demand-side investment creates equivalent new work. The WWII manufacturing analogy is misleading. Federal investment in physical infrastructure created more physical work because the cognitive tools were human brains. Federal investment in AI infrastructure creates less human cognitive work because the cognitive tools are now software. The multiplier is inverted.

3. AI will create "new lines of work" comparable to historical new work. This is asserted, not argued. Historical new work added net cognitive capacity to the economy. AI new work may replace cognitive capacity faster than it creates new cognitive niches.

4. "Young and educated workers" will absorb whatever new work AI generates. Autor's finding that new work skews young and educated is presented as reassuring. But what if the population of "young, educated, urban" workers already exceeds the new work AI will create? The premium doesn't survive oversupply.

5. Autor's own caution about "too soon to tell" is treated as genuine uncertainty rather than methodological cowardice. The evidence from AI deployment already available in 2026—coding, legal research, diagnostics, content—suggests a pattern that is definitively not "too soon to tell." It is clearly not following the historical template.

SOCIAL FUNCTION

Transition Management / Soft Landing Theater. This article performs the function of making the displacement discourse feel like it is being taken seriously at an elite level—MIT, peer-reviewed research, measured nuance—when in fact it is providing intellectual cover for the same "trust the process" framing that has characterized every prior technology transition narrative. The study is real. The conclusion is not actionable. The article publishes it anyway because publishing it satisfies the ritual requirement of engagement without requiring anyone to state the structural conclusion.

THE VERDICT

This article is a well-sourced deferral mechanism. It performs scholarship while avoiding the central question: that AI is not a new technology applying historical displacement patterns, but a structural discontinuity that breaks the underlying mechanism by which previous technologies always ultimately created more human-valuable work.

Autor's own data—that new work erodes, that automation consumes it, that scarcity is transient—is the obituary the article refuses to print. It publishes the biography instead.