Designing AI and Adapting Society to Optimize for Child Development: Lessons from 200 Years of Technological Progress

When steam power arrived in 1770s, society spent decades training workers to use new machines. But few asked a fundamental question about kids: how do we design new technology and organize society so that children thrive when it is adopted? Serious consideration of that question came 60-100 years later (too late to minimize harms or to optimize benefits). Are we repeating this with AI? Millions are being spent on AI literacy for students and getting AI tools into classrooms (and in front of kids outside of schools) without first asking whether and how AI affects broad child development (cognition, executive function, identity, agency, attachment, play). We have an opportunity to ask (and answer) this question now, while we can still shape technological design and societal response. This framework draws on 200+ years of lessons from societal adaptation to Industrial Revolutions (IRs) to consider how we design AI and adapt as a society to optimize for child development. The core insight: societies took 20-100 years to adapt to past IRs and optimize for children's development. We can (and probably must) compress this to 5-10 years by learning from what worked (and what didn't).

The Unbundling of American Childhood

Childhood is unbundling in ways that create both opportunity and risk. While real benefits are emerging, we're also starving critical developmental inputs in the process. This document explores what's being lost and gained, why it matters, and what to do about it.