Gopnikism

Last updated: Tuesday, 02 July 2024

“Gopnikism” refers to the belief that human cognition, particularly in children, is fundamentally exploratory and imaginative, driven by a natural curiosity and capacity for learning through play and experimentation.

After cognitive scientist and developmental psychologist Alison Gopnik, whose research shows how children learn about the world by exploring it, using a hypothesis-testing approach similar to scientific inquiry. Children are adept at understanding cause-and-effect relationships, forming and testing theories about how the world works.

(Not just putting cameras on babies.)

Implications for Generative AI

Gopnikism argues that generative AI models should be viewed not as intelligent agents but rather as “cultural technologies” – tools that enhance cultural transmission, serving as efficient “imitation engines”.

A very common trope is to treat LLMs as if they were intelligent agents going out in the world and doing things. That’s just a category mistake. A much better way of thinking about them is as a technology that allows humans to access information from many other humans and use that information to make decisions. We have been doing this for as long as we’ve been human. — Alison Gopnik, in “How to Raise Your Artificial Intelligence” (2024)

LLMs and generative AI models may excel at transmitting and replicating existing information from their training data, but they will always struggle with innovation, exploration, and causal learning – abilities that young children possess in abundance.

This understanding can inform appropriate use cases and boundaries for generative AI.

The goal should not be to replicate human cognition, but to build systems that can more flexibly learn, explore, and test ideas. Crucially, these systems should be viewed as tools to scaffold or augment human creativity and problem-solving, not as autonomous agents.

  • [?] How can we combine the transmission capabilities of LLMs with more flexible, exploratory learning?
  • [?] How might AI-lubricated cultural transmission impact diversity of thought and innovation?
  • [?] What are the implications of Gopnikism for human-computer interaction?
  • [?] What do Gopnikist models of cultural transmission imply for retrieval-augmented generation (rag)?