The Decoupling of Output and Consumption
A booming Gross Domestic Product (GDP) has historically been the ultimate indicator of a thriving society, operating on the established macroeconomic assumption that a rising tide lifts all boats. However, the intelligence displacement crisis introduces a highly unexpected and paradoxical reality: a booming economy that completely decouples from human prosperity. Major financial institutions, such as Goldman Sachs, project that generative AI could raise global GDP by $7 trillion over the next decade. Yet, the Citrini scenario warns that this massive nominal productivity will manifest as “Ghost GDP”.
Ghost GDP represents economic output that technically shows up in national accounts and corporate earnings but entirely fails to circulate through the real economy. For generations, economists dismissed the “Luddite Fallacy,” arguing that while automation destroys specific jobs, it ultimately lowers prices and creates new industries, allowing human labor to transition and continue consuming. But autonomous agentic AI breaks this circular flow of income. An AI agent does not sleep, eat, travel, buy houses, or purchase discretionary goods. Therefore, while AI can generate massive economic value by executing complex workflows autonomously, the velocity of the money it generates is essentially zero in the real human economy. The output exists purely in the digital realm, leading to a localized depression in consumer discretionary sectors even as macroeconomic headlines boast of historic, record-breaking growth.
The Marginal Propensity to Consume (MPC) Divide
To understand exactly why this decoupling is so dangerous, one must look at a foundational economic metric: the Marginal Propensity to Consume (MPC). MPC measures how much of an additional dollar of wealth a demographic will spend rather than save. Extensive research utilizing conjoint distributions of income and wealth reveals a stark divide. The overall MPC in the economy is roughly 0.10, but it is deeply stratified by wealth: the lowest wealth quintile has an MPC of 0.15, meaning they spend a significant portion of their earnings out of necessity. Conversely, the highest wealth quintile has an MPC of just 0.06, tending to hoard additional income in financial assets rather than cycling it back into the consumer economy.
Generative AI introduces a terrifying new actor into this equation: a highly productive digital worker with an MPC of exactly zero. When artificial intelligence automates the cognitive tasks of a middle-class professional, the wages that would have been spent on local restaurants, retail, and services are instantly converted into corporate profits. This dynamic evokes deep alienation and frustration among the working and middle classes. They are forced to live in an economy where they are constantly bombarded with news of skyrocketing tech valuations and AI-driven efficiency gains, yet they find themselves struggling with stagnant wages and an increasingly fragile job market.
Capital Concentration
If the immense economic surplus generated by AI is not flowing to human workers, where is it going? The answer is unprecedented capital concentration. The wealth bypasses the traditional labor market and is routed directly into massive capital expenditures required to sustain digital infrastructure. To meet the unquenchable demand for AI processing loads, global data centers are projected to require a staggering $6.7 trillion in capital outlays by 2030.
Consider the story of a fully automated corporate accounting firm. A decade ago, this firm might have employed 500 mid-level accountants earning a collective $50 million annually. Those salaries formed the bedrock of the local community, sustaining car dealerships, funding public schools through property taxes, and driving local retail. Today, the firm operates with just ten human partners and a massive AI server cluster. Because of AI’s incredible processing speed, the firm’s revenue has doubled to $100 million. On paper, GDP has exploded. However, the $50 million payroll that used to circulate through the local town is gone. Instead, that capital is routed directly to Silicon Valley hyperscalers to pay for cloud compute, and the surplus is distributed as dividends to a concentrated pool of wealthy shareholders.
This leads to dangerous “circular financing” loops, where tech giants invest heavily in each other’s infrastructure - artificially bidding up stock prices without creating broad-based economic growth. The result is an intensely fragile, K-shaped economy. The upper arm of the “K” continues to thrive on liquid asset appreciation and corporate profits, while the lower arm - the 80% of Americans who rely entirely on labor income and have minimal financial reserves - faces deteriorating conditions as the capital they rely on is hoarded inside the walls of remote data centers.