S5 Ep40: Our workless future

Episode Summary

Executive Summary: The episode explores Anton Korinek’s case that advanced AI could eventually make human labor economically redundant, creating a future where work is scarce but output is abundant. The discussion focuses on who would still work, how incomes could be supported through UBI or shared capital ownership, and what tax base and institutions could fund redistribution without distorting innovation.

Main Topics: AI as a potentially labor-replacing general technology (Priority: 5/5): Korinek argues AI may differ from past innovations because it could eventually match human cognitive capabilities across many tasks, making human labor redundant in a way previous technologies did not. Timing and uncertainty around labor redundancy (Priority: 4/5): The conversation emphasizes that no one knows when this transition arrives, but AI researchers commonly place it in the mid-21st century, with some extreme forecasts much earlier. Which workers would remain employed (Priority: 4/5): If labor demand falls, the market would likely keep the most productive workers and those who genuinely enjoy working, while also preserving freedom for people to keep working even if AI is better. Income support and universal basic income (Priority: 5/5): A major challenge is how people will survive if they are not working. Korinek argues societies may need institutions such as UBI or similar benefits that scale up as labor demand declines. Taxation in a workless economy (Priority: 5/5): The discussion examines how to fund redistribution when labor income shrinks, highlighting taxes on land, pollution, monopoly rents, and other fixed-supply or rent-based bases as relatively efficient options. Inequality, social value of work, and global justice (Priority: 4/5): The episode notes that workless futures could increase inequality through capital ownership and raise difficult questions about whether redistribution can extend beyond rich countries to the global south. Alternative mechanisms for sharing AI gains (Priority: 3/5): Beyond UBI, Korinek suggests broader capital ownership and even billionaire philanthropy as possible ways to distribute AI-driven gains more broadly.

Key Arguments: AI may eventually become capable of performing most or all economically valuable human tasks because the human brain is itself a computational system and machines are improving rapidly. Past technological change improved living standards overall, but it also produced winners and losers; AI could be more disruptive than industrial-era technologies because it may substitute for cognition itself. The likely transition is not immediate, but economists should prepare now because AI researchers consider human-level capability plausible within this century. Even in a labor-scarce future, people should retain freedom to work; the market would simply pay little if AI can do the job better. Without new institutions, a sharp fall in labor demand would create mass misery, so income support systems must be designed in advance. UBI or similar transfers could be financed by taxing land, pollution, and monopoly rents, which are less distortionary than taxing labor in a machine-driven economy. A future without work may shift inequality from wage differences to capital ownership, inheritances, and entrepreneurship gains. Global redistribution matters: if advanced economies automate first, poorer countries could be left behind unless institutions extend support internationally. An alternative to pure UBI is broader and more equitable ownership of AI-related capital, or large-scale philanthropy from AI-owning elites.

Data Points: Historical window discussed: Past 250 years - Korinek contrasts the post-Industrial Revolution era with prior subsistence conditions to argue technology has usually helped humanity overall. Near-term AI forecast cited: 2029 - Mentioned as an extreme prediction from Elon Musk for when human-level AI might arrive. Mid-century AI forecast cited: 2050 - One of the timelines discussed from surveys of AI researchers. Alternative AI forecast cited: 2065 - Another survey-based estimate for when advanced AI could reach human-like capability. Current labor market condition: Shortages everywhere - Korinek notes that in advanced countries labor demand is currently very high, unlike the future scenario being discussed. Paper identifier: Discussion paper 17438 - The CEPR paper being discussed is identified by number near the end of the episode.

Pivotal Quotes: "we may be economically redundant" — Anton Korinek: The interviewer highlights this phrase as the central concern of the paper. "if the artificial intelligence can do it so much better than you, you will earn only a pittance" — Anton Korinek: Korinek explains that people could still work by choice even if AI outcompetes them, but wages would be very low. "we really need to prepare to have institutions that will provide for people's essential needs" — Anton Korinek: Korinek describes the policy challenge of preventing mass misery in a future with sharply reduced labor demand.

Implications: The episode argues policymakers should treat AI-driven labor displacement as a serious planning issue now, not a distant fantasy. Future stability may depend on new redistribution systems, broader capital ownership, and tax reform before labor income erodes.