The attention economy is not a metaphor. It is a production system in which human cognitive engagement — the directed focus of minds — functions as the primary raw material for the generation of surplus value. When billions of people scroll, click, pause, react, share, and watch, they are not merely consuming entertainment or information. They are performing unpaid labor: producing behavioral data, generating advertising impressions, training recommendation algorithms, and validating content signals that are then sold to third parties. The product is not the feed. The product is the attention, packaged and resold at scale.
This reframing matters because it relocates the attention economy within the classical tradition of labor analysis. Adam Smith identified the division of labor as the engine of productive efficiency. Karl Marx showed how surplus value is extracted when workers produce more than they are compensated for. The attention economy reproduces this logic in digital space. Users produce measurable economic value — advertising revenue, platform valuation, data assets — without receiving wages, benefits, or ownership stakes. The compensation offered is access to the platform itself: a form of payment-in-kind that conceals the asymmetry beneath a veil of free service.
At the collective scale, this arrangement produces several systemic distortions. First, it concentrates wealth in the hands of platform owners at a rate disconnected from any traditional metric of labor contribution. Meta, Alphabet, and ByteDance generate hundreds of billions in annual revenue by mobilizing the attention of populations who receive none of that revenue. Second, it restructures the relationship between work and leisure. When a worker scrolls Instagram during their commute, they are simultaneously off the clock from their employer and fully on the clock for the platform. The boundary between productive time and free time dissolves. Third, it creates a feedback loop between extraction and optimization: platforms are incentivized to maximize attention capture, which drives the deployment of ever more sophisticated behavioral engineering — notifications, variable reward schedules, social comparison triggers — which further normalizes continuous engagement as the default condition of social participation.
The political economy of attention extraction also intersects with inequality. Platforms extract disproportionately from populations with less economic power — younger users, users in the Global South, users with fewer alternative means of entertainment or information — because these populations tend to have more time to give and less capacity to opt out. Attention extraction is thus not distributionally neutral. It is a mechanism that transfers value from the relatively dispossessed to the relatively concentrated.
Critics from the left argue this constitutes a new form of primitive accumulation: the enclosure of cognitive commons. The mental bandwidth of entire populations is being enclosed, standardized, and monetized without consent beyond the fine print of terms of service agreements that no population could meaningfully negotiate. Critics from the right argue instead that users voluntarily exchange attention for valuable services, and that any coercion narrative misreads free choice. Both positions miss the structural point: when the entire information environment is designed by parties with a direct financial interest in maximizing your engagement, the conditions for meaningful voluntary choice are systematically undermined.
Understanding the attention economy as labor extraction has implications for policy, organizing, and design. If attention constitutes labor, then data ownership frameworks, revenue-sharing proposals, platform cooperative models, and attention unions all become coherent responses rather than utopian fantasies. The question shifts from "how do we use technology responsibly" to "who owns the product of cognitive work, and who decides how it is organized." That is a labor question. Answering it requires labor thinking.