CONCEPT
The Ecological Model of AI
Neal Stephenson's framework for thinking about artificial intelligence not as a monolithic technology requiring a single policy response but as a diverse population of non-human intelligences, each occupying a distinct ecological niche with distinct capabilities, risks, and governance requirements.
The ecological model is the central analytical contribution of
Neal Stephenson's 2025 essay “Remarks on AI from NZ,” and it offers a direct remedy for the characteristic failure of public AI discourse: the demand for a single response to a phenomenon that is, in practice, as diverse as the animal kingdom. Stephenson observes that humanity has always coexisted with non-human intelligences—animals—and that the rich taxonomic tradition of understanding distinct species in distinct niches provides a more useful model for thinking about AI than the binary of salvation-or-catastrophe that dominates policy debate. On his taxonomy, conversational AIs like ChatGPT are lapdogs: “acutely tuned in to humans and basically exist to make life easier for us.” Specialized task-oriented AIs are sheepdogs, doing useful things humans cannot do themselves, and in Stephenson's view the most consequential in the long run. Narrow AIs excellent at specific tasks but oblivious to human concerns are dragonflies. AIs aware of