Will the AI economy create a permanent underclass? | Kenneth Rogoff - The Guardian

TEXT ANALYSIS: Rogoff's Guardian Piece

The Dissection

This is elite anxiety theater—performed concern by a credentialed insider who sees the shape of the crisis but cannot name its mechanism. Rogoff correctly identifies that AI threatens mass displacement, that developing nations face a different and worse calculus, and that the current distribution architecture is inadequate. He then retreats into the comfortable vocabulary of mainstream economics: redistribute gains, expand safety nets, invest in human capital. The diagnosis is partly accurate. The prescription is delusion dressed in academic prose.

The Core Fallacy

Treating structural extinction as a distribution problem.

Rogoff writes as though the solution to AI displacement is "finding ways to distribute the benefits more broadly" or praying for "generous universal basic income." This is the canonical institutional blind spot. DT does not predict that redistribution will fail because elites are wicked or politics is broken—though both are true. DT predicts redistribution fails because the mechanism linking productivity to wages to consumption is being severed. You cannot redistribute what no longer flows through the human wage circuit. UBI funded by AI profits is a transfer payment, not productive participation. It may preserve caloric consumption. It cannot preserve the post-WWII social contract that required mass employment as its connective tissue.

Hidden Assumptions

1. Participation remains possible. The article assumes displaced workers can be redirected into economically meaningful roles with sufficient retraining or redistribution. DT says this is increasingly false at scale.
2. The nation-state is the relevant unit. Rogoff frames this as a competition between countries—"winning" the AI race, carving out a place in the supply chain. But DT's structural argument operates at the level of the global production function. Whether you are a "winner" or "loser" in the geopolitical AI race, your domestic mass employment base faces the same compression.
3. Institutional responses are adequate given sufficient will. The article treats political dysfunction as the main obstacle to managing AI's impact. DT treats political dysfunction as a symptom of structural transformation that will worsen, not improve, as the productive participation circuit collapses.
4. India's brain drain is a solvable policy problem. The framing treats talent export as a governance failure rather than an equilibrium response to the fact that human cognitive labor in India has been serving demand generated elsewhere—demand that DT predicts will evaporate.

Social Function

Prestige signaling with progressive packaging. This is a credentialed economist performing the expected progressive concern for global inequality while remaining institutionally safe. It will be shared by people who want to appear thoughtful about technology's impact. It generates no discomfort. It proposes nothing that threatens existing power structures. It will age poorly in exact proportion to its popularity.

The Verdict

Rogoff correctly identifies the phenomenon and then immediately retreats into framework that cannot process it. The piece is intellectually honest about symptoms and professionally dishonest about prognosis. "No one really knows what such a world would look like" is a confession of theoretical bankruptcy, not epistemic humility. DT knows what that world looks like: not post-scarcity, not dystopia, but managed decline for the majority with islands of Sovereign wealth and a vast Serf class sustained by transfer payments the political will to maintain will erode as the tax base hollows. The article performs concern for the global south while ignoring that the "winners" face the same structural rupture internally—just on a longer fuse.