The AI Serpent

In the AI Story

The displacement phase is already underway, absorbing the most automatable tasks and producing income shocks for workers concentrated in them. Entry-level legal, consulting, and accounting positions are contracting. Freelance markets for standard copywriting, basic design, and routine code production are shrinking. Layoffs at technology companies cite AI efficiency as justification. The displacement phase affects the margin of the professional middle class, not its center, but the margin is wider than aggregate employment statistics suggest and is where the serpent enters.

The compression phase — the mechanism described under competitive compression — extends distributional impact from directly displaced workers to the broader professional middle class. This phase is more insidious than displacement precisely because it is invisible to standard measurement. Workers remain employed; aggregate statistics show a functioning labor market; distributional reality shows the erosion of the economic position of the professional middle class relative to populations pulling away above them.

The exclusion phase affects primarily the global poor, widening the gap between populations that participate in the AI economy and populations that do not. This is not active deprivation but differential acceleration: AI-augmented populations accelerate while excluded populations remain at their prior trajectory. The gap compounds over time, and compounding makes future catch-up progressively more difficult because the institutional infrastructure needed to participate in the AI economy becomes more complex and expensive as the AI economy advances.

Each phase has a tipping point beyond which distributional dynamics become self-reinforcing. Displacement tips when the number of displaced workers exceeds retraining and social insurance capacity. Compression tips when middle-class wage erosion reduces the tax base that funds public investments moderating compression. Exclusion tips when the gap between AI and non-AI economies exceeds the political will of populations asked to fund the institutional infrastructure bridging it. The political dynamics of populations experiencing absolute decline differ qualitatively from the dynamics of populations experiencing relative stagnation: historical serpent phases have produced political upheaval threatening institutional order itself, not merely dissatisfaction channeled through existing institutions.

Origin

The serpent concept is an analytical extension of Milanovic's elephant framework, applying the same distributional logic to a scenario where institutional moderation fails. The historical reference is the well-documented immiseration of specific populations during early industrialization — the handloom weavers being the canonical example — whose conditions deteriorated in absolute terms for decades before institutional responses eventually moderated the broader industrial revolution. The serpent names the risk that AI could produce a similar phase for the professional middle class at compressed timescale.

Key Ideas

Relative stagnation versus absolute decline. The elephant shows the first; the serpent shows the second. The transition from one to the other occurs when the mechanisms of compression persist without institutional moderation.

Three phases, each with a tipping point. Displacement tips when transition capacity is exceeded; compression tips when tax-base erosion undermines moderating institutions; exclusion tips when the AI-economy gap exceeds bridging will. Each tipping point is reachable within the current trajectory.

Political consequences differ in kind. Relative stagnation produces political dissatisfaction; absolute decline produces political upheaval that threatens institutional order itself. The serpent phase operates outside the bounds of normal institutional politics.

Speed amplifies severity. Historical serpent phases unfolded over decades; the AI serpent could unfold in years. Speed compresses institutional response capacity to a point where moderating interventions arrive too late to prevent the political upheaval that their absence will produce.

The serpent is conditional, not inevitable. Each phase can be moderated by specific institutional interventions — robust social insurance, strengthened labor bargaining, portable benefits, international coordination. The moderation is technically feasible and politically contested.

Debates & Critiques

The most serious objection is that the AI transition may not produce displacement and compression at the scale historical analogies suggest, because AI tools may create new tasks and occupations at rates that offset the displacement of old ones. Milanovic's response is that the question is not whether net displacement is positive or negative in aggregate but whether the specific populations bearing the displacement costs have the institutional support to transition; without the support, even a net-positive transition produces a serpent phase for the specific populations who lose in the reshuffling.