You Are an Unpaid Worker — and the World’s Accounting Systems Can’t See It
When you scroll a social media feed, you are simultaneously producing a distinct economic output — behavioral data that trains AI systems and generates hundreds of billions in revenue. This production is uncompensated and invisible to GDP.
What GDP Misses
The System of National Accounts classifies your screen time as idle leisure. The firms monetizing it disagree.
Screen Time = “Leisure”
Every statistical agency on Earth classifies digital engagement as non-market leisure consumption. You’re scrolling in a park, economically speaking.
$1.3 Trillion in Monetized Data
Digital advertising ($680B), enterprise AI ($300–500B), and data brokerage ($250–350B) are all built on behavioral data generated during this “leisure.”
The Measurement Blind Spot
No minimum wage law covers this labor. No union bargains for it. The market that would constrain extraction does not exist.
Entangled Time: A Third Kind of Time
Since 1965, economics has partitioned time into two states: work and leisure. The digital economy created a third.
Becker’s Two-State Partition
Gary Becker (1965): all time is either work (selling time for wages) or consumption (enjoying goods). These states are mutually exclusive. GDP is built on this partition.
T̄ = Tw + Tc where Tw ∩ Tc = ∅
Simultaneous Production-Consumption
Digital engagement is both consumption (you enjoy the content) and production (your behavioral data trains AI systems). This Entangled Time is the missing third state.
Tw + Tp + Te + Bmin = T̄
The Autocatalytic Loop
Zero-cost data labor → firms reinvest in better algorithms → higher engagement → more data → the loop closes. The firm’s capacity is ultimately bounded by the total waking hours of the human species.
Kdss = λN(T̄ − Bmin) / δd The Numbers
Calibrated estimates from publicly available data (2024–2025).
hours/year
Unpaid Data Labor
5.5 billion connected people × 6.5 hours/day of screen time = 13 trillion hours of Entangled Time per year. That is 4× the total formal labor force.
per year
Conservative Dark GDP
Using only directly observable revenue channels (advertising, enterprise AI, data brokerage). The implied shadow wage: roughly $0.10/hour — below every minimum wage on Earth.
adjusted estimate
Full Measurement Correction
Including unmeasured channels (foundation model training, implicit RLHF, background data collection, biometric bandwidth), the adjusted estimate recovers the majority of the global labor share decline.
optimal wage
Zero-Wage Equilibrium
The profit-maximizing fiat wage for data labor is exactly zero — not from greed, but from market structure. The labor supply is perfectly inelastic at the biological maximum. The firm cannot increase data by paying more.
Three Institutional Redesigns
Not utopian fantasies — specific policy instruments grounded in the model’s equilibrium structure, each with precedent in existing regulatory frameworks.
Algorithmic Monopsony Standard
Replace the Consumer Welfare Standard — which cannot address zero-price platforms — with an upstream focus: “Is the data laborer being underpaid?” instead of “Is the consumer being overcharged?”
Algorithmic Severance Tax
A Pigouvian levy on extracted data capital, calibrated to the shadow wage. Because labor supply is nearly perfectly inelastic, deadweight loss is correspondingly modest — a rare case of an efficient Pigouvian tax.
Cognitive Depreciation Allowance
Explicit compensation for the endogenous depreciation of autonomous cognitive capacity caused by sustained interface use — funded by Severance Tax revenue. Not UBI. A depreciation allowance for human capital.
The window is closing. The model proves these interventions are effective only during the transition — before cognitive capital depreciates irreversibly. After the transition, the equilibrium is a trap. Each year of inaction narrows the window.
Read the Research
“The Economics of Entangled Time: Simultaneous Production-Consumption, Dark GDP, and the Macroeconomic Architecture of Algorithmic Monopsony”
Working paper · March 2026 · Nav Vaidhyanathan
Discuss This Research
Interested in the policy implications of Dark GDP? Exploring the economics of digital labor for your organization? Let's talk.