The Useless Class

In the AI Story

The mechanism generating uselessness operates through what might be called the 'good enough threshold.' Before AI, expertise commanded a premium because it produced outputs of superior quality. A senior developer's code was cleaner, more robust, more maintainable than a junior's—a quality gap justifying salary differentials of two or three times. AI compresses this gap not by making senior work less valuable in absolute terms but by raising the floor of junior-plus-AI capability. When an AI-assisted junior produces code eighty percent as good as an unassisted senior, the market asks whether the remaining twenty percent justifies the salary differential. Often the answer is no—eighty percent quality at forty percent cost is, for most commercial purposes, economically superior. The senior's advantage remains real. It stops being worth paying for. This is not unemployment but devaluation—still working, still producing, but producing output the market values less with each AI capability improvement.

The psychological dimension distinguishes the useless class from traditional unemployment. Unemployment provides a clear narrative: I lost my job because the factory closed, the company downsized, the economy crashed. Uselessness provides no such clarity. You still have a job. You just can't explain why it pays less each year, why the sense of mastery has been replaced by vague fungibility, why the expertise you spent decades building feels less relevant each quarter. The identity crisis is more corrosive than economic hardship because it is existential. Modern humans derive purpose, standing, and self-understanding from what they do. 'What do you do?' is synonymous with 'Who are you?' When the answer becomes 'I do what a machine does, only slower and costlier,' the identity the answer was supposed to provide collapses. You exist. You have capacities. But the capacities around which you built your sense of self are no longer scarce, and scarcity—not quality—is what markets reward.

Harari has been frank about the inadequacy of standard reassurances. The claim that new technologies always create new jobs is historically true in aggregate but misleading about transitions. Handloom weavers displaced in the 1810s did not become factory managers; some of their grandchildren did, after decades of poverty and upheaval. The aggregate balanced eventually. The transition generation bore the cost. And the AI transition compresses displacement timelines (months rather than decades) while offering no evidence that job-creation timelines have similarly compressed. The deeper challenge is that uselessness is not merely economic but philosophical. If human purpose has been defined by productive labor for centuries, and if AI makes that labor economically redundant, then the species faces not a jobs crisis but a meaning crisis—a crisis for which the standard economic solutions (retraining, universal basic income, job guarantees) are necessary but insufficient because they address subsistence without addressing significance.

Origin

The 'useless class' concept first appeared in Homo Deus (2015) and was elaborated in 21 Lessons for the 21st Century (2018), where Harari devoted an entire chapter to the problem. The term generated immediate controversy—critics accused Harari of dehumanization, arguing that no person is 'useless' and that the framing itself enacts the utilitarian reduction he purports to critique. Harari's response has been that the term describes not an ontological status but an economic classification—how market societies categorize people—and that refusing to name the category does not prevent its formation but merely makes it harder to address.

The concept synthesizes technological unemployment theory (Keynes's 1930 'Economic Possibilities for our Grandchildren'), observations about automation's uneven impact across skill levels (Autor, Levy, Murnane on routine cognitive work), and Harari's distinctive attention to the identity-constitutive role of labor in modernity. His claim is not that uselessness is inevitable but that it is the default trajectory absent deliberate institutional intervention—and that the speed of AI development is outpacing the speed of institutional response by orders of magnitude.

Key Ideas

Economic irrelevance, not oppression. The useless class is not exploited for labor or surplus but simply unneeded—a status more psychologically devastating than traditional forms of economic subjugation.

The 'good enough' threshold compresses expertise premiums. AI raises the floor of competent performance, making depth less economically valuable even when depth remains epistemologically real—the market stops paying for quality it no longer needs.

Identity crisis exceeds economic hardship. Modern humans derive purpose from productive contribution; when that contribution becomes redundant, the loss is existential—affecting self-understanding, social standing, narrative coherence.

Transition speed outpaces adaptation. Historical technology transitions offered decades for job creation to replace displacement; AI compresses displacement to months while offering no evidence creation timelines have similarly accelerated.

Meaning crisis, not merely jobs crisis. Retraining and income support address subsistence but not significance—the deeper problem is civilizational loss of the shared fiction that labor confers purpose and worth.