Labor and Artificial Intelligence

Mistakes are inevitable, but fortunately human mistakes are typically heterogenous. Using the same machine learning model for high stakes decisions creates consistency while amplifying the weaknesses, biases, and idiosyncrasies of the original model. When the same person re-encounters the same model or models trained on the same dataset, she might be wrongly rejected again and again. Thus algorithmic monoculture could lead to consistent ill-treatment of individual people by homogenizing the decision outcomes they experience. Is it wrong to allow the quirks (...) of an algorithmic system to consistently exclude a small number of people from consequential opportunities? This paper will argue that consistent and arbitrary exclusion is often unfair, even when it is divorced from group-based bias or discrimination. (shrink)

High-income societies are undergoing a historic inversion: population is changing from strategic asset to thermodynamic liability. Declining fossil-fuel EROI, accelerating automation, and the fiscal collapse of labour-based states are making demographic contraction the inevitable correction of a system that over-invested in people during a brief energy windfall. This essay outlines the institutional form suited to an era of biophysical limits and machine abundance: the Sovereign Neganthropic Economy. It advocates deliberate transition toward smaller, debt-free, high-technology polities that publicly own robotic and (...) algorithmic capital, issue money against energy surplus rather than future growth, and redirect human time from wage-labour to entropy-reducing activities (care, research, restoration). The argument confronts the political agency required, the uncertainty of full automation, and the dangerous global consequences of unmanaged contraction in the ageing Core. (shrink)

Economic change can bring many benefits, but it can also upset the economic positions and prospects of individuals and communities by dramatically curtailing access to decent employment. This paper explores the question of how a society ought to address those left behind by economic change. The paper’s first goal is to defend the claim that the best way to frame the issue of access to decent employment is not just as an issue of economic policy but also as a public (...) health issue, a framing that it justifies, in part, through the case study of deaths of despair, a phenomenon famously explored by Anne Case and Angus Deaton. This framing clarifies the moral urgency of expanding access to decent employment. The paper’s second goal is to discuss the values at stake in the problem and in the possible solutions. I argue that 1) the state should implement policies friendlier to private sector employment and 2) the state should serve as an employer of last resort, at least when the work created is productive and doesn’t create out-of-control inflationary pressures. (shrink)

Work is often said to be hard to define. A precise working definition may nevertheless be valuable for analytical purposes, such as discussing justice in the distribution of work or the future of work. This paper takes a conceptual engineering approach to the concept of ‘work’. It examines the most common features of definitions of work in the contemporary philosophy of work: pay, negation of leisure, effort, social contribution, necessity/instrumentality and production of a benefit/external good. Of these, it argues that (...) the latter two can be refined into two equally suitable, co-extensive working definitions: Work is activity that is instrumental and intermediary to satisfy needs and wants, and Work is activity that produces a benefit (or intermediary product to a benefit) external to the worker(s) doing the activity. These are useful working definitions by virtue of coming close to a core of ordinary language ‘work’, being easily adaptable for various purposes, and for tracking a central feature of the human condition of particular importance for the philosophy of work. (shrink)

This study examines the impact of demographic shifts and broader macroeconomic factors on labor productivity in 23 European Union countries (EU-23) over the period 2005–2022. The analysis employs dynamic panel estimation techniques, including the Generalized Method of Moments (GMM), Random Effects (RE), and Autoregressive Distributed Lag (ARDL) models, to investigate the effects of population aging, unemployment, inflation, GDP per capita, research and development (R&D), life expectancy, and capital formation on productivity dynamics. -/- Unit root and cointegration tests confirm the existence (...) of long-run relationships among variables with different levels of integration. The findings indicate that inflation exerts a significant negative effect on productivity, whereas GDP per capita and R&D investment consistently enhance efficiency and output per worker. Life expectancy appears to have a negative impact, likely reflecting demographic pressures, while unemployment shows an unexpected positive association with productivity – possibly due to structural changes in the labor market and the adoption of new technologies. -/- The ARDL error-correction model confirms long-run convergence, with persistent effects of productivity determinants over time. Overall, the results suggest that sustaining productivity growth in Europe depends on macroeconomic stability, innovation, human capital accumulation, and demographic adaptation. (shrink)

This essay originally appeared as a guest post on Shtetl-Optimized (see link).

Today, generative artificial intelligence systems have become widespread and are increasingly used for commercial purposes. Since these systems are trained on copyrighted content, this raises the issue of exploitation of creative labor. John Rawls’s theory of justice offers a promising solution to this problem. Because of its priority of basic rights over considerations of economic gain, this theory of justice imposes strict restrictions on the use of other people’s content in training neural networks, and therefore requires government regulation in this (...) area. In addition, Rawls’s theory also justifies the provision of a guaranteed basic income both as a means of compensation for past violations of rights and as a means of distributing fair shares of social wealth. (shrink)

How can work be a genuine good in life? I argue that this requires overcoming a problem akin to that studied by Marx scholars as the problem of work, freedom and necessity: how can work be something we genuinely want to do, given that its content is not up to us, but is determined by necessity? I argue that the answer involves valuing contributing to the good of others, typically as valuing active pro-sociality – that is, valuing actively doing something (...) good for others. This makes work better in one way, and may even make work something we are genuinely glad to have in our lives. Contemporary philosophical thinking about good work tends to focus on how work can be good for the person doing it, by providing, for example, self-realization or social relationships, while underappreciating the special importance of valuing social contribution. People will typically only really want work if they want a part of their lives to be about the good of others. This also means that work may be a part of the best life, something we should take into account when discussing work-related policies and the desirability of a ‘post-work’ future. (shrink)

This paper envisions a hybrid society where humans, robots, and AI agents coexist as intelligent, self-interested entities capable of planning, reasoning, acting, collaborating, competing, and evolving. Drawing on sociological theories (structural functionalism, conflict theory, exchange theory, constructivism), systems theory, economics, psychology, ethics, and management science, we analyze emergent societal structures and interactions. We consider AI traits like vast knowledge, continuous operation, rapid replication, and instant creation. Economics examines resource constraints, such as energy, and disparities in AI resource control. Psychology explores (...) AI behaviors resembling selfishness or tribalism. Ethics addresses equality and moral obligations among humans and AI. Management science investigates coordination and conflict resolution. Systems theory models this society as a complex adaptive system, emphasizing openness, self-organization, and interconnectedness. We propose AI design principles to ensure adaptability, ethical alignment, collaboration, resilience, and systemic integration, fostering a harmonious and innovative society. (shrink)

As frontier AI systems become more powerful and general, governments face new distributional and fiscal questions which are often discussed with more confidence than rigor. This paper uses standard tools from political philosophy and public economics to draw five lessons for taxing AI and for economic distribution in an AGI era. First, policy proposals must tackle the standard equity-efficiency tradeoff, persistent disagreement about what an equitable society looks like, and the difficulties of predicting and assessing the distributional effects of new (...) policies. Second, advanced AI is unlike ordinary capital: It relies directly on centuries of diffuse knowledge and data and is, in a unique and strong sense, the product of humanity. One consequence is that AI companies have a moral claim on a relatively small share of the income their models create, as can be illustrated through the concept of Shapley value. Third, the economic abilities of AI systems are likely more precisely measurable than those of their human counterparts, which raises the possibility of low-distortion AI taxes within the theory of optimal taxation. Fourth, beyond equity, there are efficiency motivations for taxing AI, including its negative externalities (e.g. pollution) and the need to finance new public goods related to AI safety, security, and governance. Fifth, while Universal Basic Income should be considered as a solution to AGI-caused labor displacement, it has significant equity and efficiency flaws and its pursuit should not come at the cost of other, perhaps equally novel, policy ideas. A more plausible solution will be morally tailored (focusing on, e.g., the needy), and will incorporate "tagging" considerations. (shrink)

This paper clarifies the concept of self-determined work as a structurally necessary condition of reality-binding in the context of advanced artificial intelligence. Building on the concept of work-integrated relational agency, which was developed in earlier works as a critique of functionalized labor, it is shown that self-determination was never meant as a normative ideal or ethical preference, but as a prerequisite for work to retain its relation to reality and its capacity for innovation. The central thesis is that work is (...) not to be understood as activity, task fulfillment, or economic function, but as a form of irreversible, world-binding action in which human Eigenzeit is not delegable. Artificial intelligence does not function here as the cause of the problem, but as a structural test case that makes visible which aspects of work can be functionally replaced—and which, in principle, cannot. While AI can formally simulate decisions, interactions, and innovation processes, it is structurally incapable of performing work-integrated relational agency, since it lacks Eigenzeit, responsibility-binding, and an irreversible relation to the world. The paper introduces the concept of Eigenzeit as a category previously missing from the debate on work and shows that self-determined work is necessary because only it can effectively bind Eigenzeit to the world. By contrast, heteronomous work reproduces functioning processes but loses its enactment of reality and increasingly generates simulation instead of innovation. The paper thus provides an ontological clarification of the concept of work under conditions of artificial simulation, without resorting to moral, psychological, or labor-market-based argumentations. Against this background, self-determined work is identified as an ontological stability condition of human world-capability in the age of AI—not as an option, but as a structural necessity. (shrink)

This paper advances the thesis that poverty is not a socio-economic condition but an ontological consequence: the result of systematically withdrawn Eigenzeit. Poverty emerges where the world can no longer be bound, where action is organized without bearing, responsibility without world, and time without irreversibility. The point of departure is the concept of Eigenzeit developed in earlier work as the non-delegable instantiation of world-time in action. Eigenzeit names the condition under which life becomes structurally viable at all, because decision irreversibly (...) binds world. The paper argues that modern systems of work, markets, and social security structurally undermine this condition. Wage labour replaces world-binding with procedural time; social security systems administer the loss of Eigenzeit without being able to remedy it. Poverty thus appears not as a lack of resources, but as a permanent impossibility of living Eigenzeit. This impossibility is conceptualized here as existential prohibition: not as repression or exclusion, but as the structural prevention of viable world-binding. Neurodivergent existences do not function as a special case in this analysis, but as early indicators. Through them, what applies more generally becomes visible: where Eigenzeit is systematically rendered impossible, it is not only individual biographies that collapse, but the ontological viability of societal orders themselves. The paper does not propose reforms; it marks a limit: poverty is not solvable as long as the world is merely administered. The analysis developed here does not arise from detached observation, but from long-term operatoric research in which Eigenzeit was systematically blocked under real institutional conditions. (shrink)