What this covers

Philosopher Jacob Howland speaks with Russ Roberts on EconTalk about the human costs of outsourcing thought and judgment to artificial intelligence. The conversation moves past familiar worries about job displacement to examine what happens to human capacities when they go unused—a concern rooted in Capacity Atrophy in Disuse. Howland argues that ceding intellectual, creative, and moral functions to AI erodes the very abilities that constitute meaningful human life, and that the technological acceleration traced by philosopher Henry Adams a century ago has only intensified. Roberts pushes back on some of these claims, particularly the notion that technological change necessarily diminishes us, while conceding Howland's deeper point about what we stand to lose.

The discussion ranges across education, where AI-generated writing threatens to short-circuit the compositional struggle through which students actually learn; reading and leisure, understood through Josef Pieper's Leisure: The Basis of Culture as receptive openness rather than idleness; and the architecture of digital life, which traps users in algorithmically tailored Plato's Cave-like chambers. The speakers also address AI's capacity to fabricate and manipulate—through deep fake technology, ideological media, and personalized psychological profile-based content—in ways that undermine our ability to judge truth. Howland's gravest concern is neither unemployment nor extinction but the surrender of human judgment and, ultimately, liberty. Roberts offers a measured counterpoint: since AI cannot be stopped, the realistic task is to consciously preserve pockets of human skill and communal experience, accepting that some atrophy is inevitable while defending what matters most.

Sharpest takeaway

Howland argues that outsourcing thinking, writing, and judgment to AI causes human capacities to atrophy and severs us from the leisure, communal experience, and openness to reality that constitute a meaningful human life—a deeper threat than job loss or extinction.

  • Human capacities atrophy when unused, so ceding intellectual, creative, and moral functions to AI erodes them
  • Leisure properly understood is receptive openness to the world, which AI's virtual realities and individualized 'caves' close off
  • AI's capacity to fabricate and manipulate threatens our ability to judge what is true, with implications for liberty and democracy

The argument · threads4 threads · 29 claims
0.85

We must consciously maintain human skills in conversation and deep engagement rather than surrendering them to technological convenience.

3 pointscentrality 5/5
  • The smartphone has degraded its user's attention span and ability to be a focused friend or family member, though new norms may yet emerge to help us deal with such technologies and maintain our humanity.

    “I see what it's done to my attention span and my ability to be a focused friend at times or family member. And, I am concerned about it.”

  • The value of reading a great work like the Odyssey is not learning what happened at the narrative level—which ChatGPT can summarize beautifully—but learning about the human experience and oneself through the arduous task of wrestling with the text.

    “You don't read the Odyssey to find out what happened...You read it to learn something about the human experience and yourself. And, that learning takes place through the arduous task of wrestling with the text.”

  • Since we cannot stop ChatGPT and AI, the right battle is to make people aware of what it may be doing to them and to deliberately decide where to maintain human skills—conversation, communication, emotional connection—whether through a Sabbath, restraints on phones, or Amish-like limits, while some atrophy is acceptable.

    “I don't think we can stop ChatGPT and AI. I think we ought to be fighting a different battle. I think we ought to be encouraging people to be aware of what it might be doing to them.”

0.75

Technological progress has reduced material hardship and created safer, more meaningful work despite nostalgic romanticization of the past.

2 pointscentrality 5/5
  • Nostalgic claims that life in 1900 was slower, deeper, and richer ignore that for most people 1900 was hard; the material and technological acceleration of the last century cost us something but also relieved widespread hardship.

    “that wasn't true for most people in 1900. Most people in 1900 had hard lives.”

  • The fear that technology will eliminate jobs is very old and has repeatedly failed to materialize; instead work has become safer and more pleasant, with dangerous farming and manufacturing jobs replaced by more meaningful skill-based work, because cheaper, resource-conserving technologies spawn new activities.

    “this is a very old worry--the worry that technology will get rid of, eliminate jobs. It hasn't.”

0.38

AI may force education to abandon rote testing and demand genuine grappling with meaning, a beneficial but uncertain shift.

1 pointcentrality 3/5
  • Because multiple-choice and regurgitation-based testing becomes pointless when ChatGPT can do it, AI may force educators to abandon parroting and ask students to grapple with what texts mean for their own lives—a change that, while uncertain, would be good for education.

    “the whole idea of multiple choice exams will be essentially impossible...we're going to force them--I think we should--to grapple with what the text means for themselves”

0.24

Hyped transformative technologies often fail to materialize, warranting skepticism about AI predictions.

1 pointcentrality 2/5
  • The driverless car, hyped roughly eight years ago as transformative, would have put millions of taxi and truck drivers out of work overnight had it become a viable technology, but it has not fulfilled its promise—illustrating that hyped AI-adjacent technologies often fail to materialize.

    “The driverless car was the rage eight or so years ago, and that was going to change the workplace--which it would have if it were viable.”