AI Cannot Produce Writing That Humans Actually Want to Read Due to Fundamental Emptiness

Source: Jay Caspian Kang. "Can A.I. Produce Writing That We Actually Want to Read? | The New Yorker." June 2, 2026. www.newyorker.com

The Gist

The author argues that while AI can copy writing styles pretty well, it can't create the kind of writing people actually enjoy reading. Even when AI gets better at mimicking famous authors, the stories it creates feel empty and lifeless because the characters don't really do anything meaningful.

Conclusion

A.I. cannot produce writing that we actually want to read because it fundamentally lacks the human capacity to create meaningful action, genuine emotion, and authentic engagement with life

Premises

  1. AI writing consistently exhibits 'tells' including formatting problems, overreliance on tortured metaphors, and characters who constantly fidget
  2. Even when AI successfully mimics stylistic elements of famous authors, the resulting passages describe absence, stasis, and emptiness rather than meaningful human experience
  3. AI-generated characters 'mostly do nothing' - they lack agency, certainty, and genuine emotional responses to their circumstances
  4. Despite technical improvements through multiple AI agents and refined instructions, the fundamental problem of lifeless, inert prose persists
  5. When tested on thousands of readers, AI writing could fool about half the people, but the most perceptive readers (10%) could still reliably detect the artificial nature
  6. The core deficiency is 'something more abstract and even a little spooky' - an absence of authentic human experience that cannot be programmed away

Assumptions

View this argument on LogicFirst.ai