Jacob Howland on the Hidden Human Costs of AI

Episode Summary

Executive Summary: Russ Roberts and Jacob Howland examine AI’s likely effects on human flourishing beyond extinction risk. Their focus is on atrophy, meaning, leisure, education, truth, and idolatry: AI may erode skills, deepen dependence, weaken reading and judgment, and tempt people to treat human-made systems as godlike authorities. Both agree AI is coming, but differ on how much loss it entails and where humans should draw boundaries.

Main Topics: AI and the loss of meaningful work (Priority: 5/5): Howland argues AI may displace jobs and remove psychologically important routines, leaving people with enforced leisure they are unprepared to use well. Roberts counters that many jobs were historically worse and that reduced work can be a gain if society adapts. Atrophy of human capacities (Priority: 5/5): The conversation centers on how outsourcing navigation, writing, reading, and judgment to AI can weaken those same abilities in humans over time. Roberts treats atrophy as sometimes acceptable; Howland sees a broader danger to character and agency. Education and reading in the age of ChatGPT (Priority: 5/5): Howland worries that students will use AI to avoid the hard work of reading and writing, producing shallow learning and less developed intellects. Roberts suggests education should shift toward deeper engagement with texts rather than recall-based testing. Leisure, receptivity, and the recovery of attention (Priority: 4/5): Drawing on Aristotle, Keynes, and Joseph Pieper, Howland argues true leisure means openness to reality, not mere idleness or screen time. Roberts agrees modern distraction undermines solitude, attention, and personal presence, and both see practices like Sabbath as protective. Truth, manipulation, and democratic vulnerability (Priority: 5/5): Howland warns that AI intensifies already serious problems of misinformation, deepfakes, surveillance, and individualized manipulation. He fears a future in which people cannot distinguish reality from algorithmically tailored persuasion. Idolatry and the danger of worshipping our creations (Priority: 4/5): Howland frames AI as a form of idolatry: a human-made system treated as if it possessed godlike powers. Roberts pushes the idea further, noting the eerie emergence of an apparently autonomous intelligence that remains built from human-generated data. AI as a technological and cultural transition (Priority: 4/5): Both see AI as unavoidable and likely to reshape norms, institutions, and habits. The question is not whether it arrives, but how humans preserve skills, judgment, and community while adapting to it.

Key Arguments: AI can be useful for rapid information gathering, drafting, and image generation, but it also tends to hallucinate, fabricate, and encourage shallow dependence. Technological acceleration is historically disruptive, and AI may accelerate that disruption even faster than earlier inventions. If machines take over more work, humans may lose one of the chief structures that gives life meaning, not merely income. Human capacities such as navigation, writing, reading, and conversational depth atrophy when they are no longer practiced. Education built around regurgitation, summaries, and multiple-choice recall will fail in an AI-rich environment; teaching must shift toward interpretation, judgment, and wrestling with texts. AI threatens to erode people’s ability to judge truth because it can tailor output to psychological profiles, amplify propaganda, and generate convincing fakes. Leisure is not just free time; it is a disciplined openness to reality, community, art, and contemplation. Sabbath-like limits on technology can help preserve human attention, moral life, and receptivity. AI resembles idolatry because it invites us to trust and revere a human-made system as if it were omniscient or omnipotent. The deepest worry is not only job loss or even extinction, but a gradual deformation of human character and social life through dependency on machine mediation.

Data Points: Podcast date: May 31st, 2023 - Opening introduction to the episode Henry Adams birth year: 1838 - Howland references Adams’ life span while discussing accelerating technological change Acceleration hypothesis period: Since around 1800 - Howland cites Adams’ view that human power has doubled roughly every decade since then Jonathan Haidt turning point: 2012 - Roberts references the year Haidt associates with worsening anxiety and mental health among children using smartphones Workweek trend: Creeping downward over a century - Roberts notes that average lifetime work hours have declined over time AI job displacement examples: Computer programming, law, education - Howland names fields he thinks AI may transform or automate heavily Safety comparison: Deaths per hundred thousands of teenagers or children in the 1970s were much higher than now - Howland uses this to illustrate how some technologies/standards have made life safer Lottery winner example: A custodian quits work, moves, buys a house - Howland uses this hypothetical to show how losing work can destroy life structure Response format of social media: Binary: thumbs up / thumbs down - Howland uses Twitter as an example of machine-like interaction conditioning AI risk estimate cited: 2% chance of destroying humanity - Roberts cites Scott Aronson’s estimate while contrasting it with a larger chance of benefit

Pivotal Quotes: "work is not particularly meaningful for a lot of people, but it's essential for their lives" — Jacob Howland: Howland distinguishes between intrinsic job satisfaction and the psychological structure work provides "the capacity to communicate to technology. We will lose the ability to express ourselves" — Russ Roberts: Roberts reframes the atrophy argument around surrendering expressive capacities to machines "I'm not interested in artificial intelligence. I'm trying to find real intelligence" — Stanley Rosen (quoted by Jacob Howland): Howland invokes his teacher to question the obsession with machine intelligence over human intelligence

Implications: Listeners are urged to treat AI as an unavoidable force that can still be bounded. The key challenge is preserving reading, writing, attention, judgment, community, and reverence for reality while deciding which skills and habits are worth defending from automation.