The 1994 Oikos paper that established ecosystem engineering as a distinct ecological category accumulated over thirty-four thousand citations and reshaped how ecology understands the relationship between organisms and environments. Haeckel's framework anticipated the category a century earlier—his insistence that organism and environment are aspects of a single relational system implied that organisms would, by their existence, alter the environments on which they depended. The beaver building its dam, the earthworm cycling soil, the human constructing cognitive infrastructure: all are ecosystem engineers, and all produce cascading effects that propagate through systems they do not directly touch. Applied to AI, the framework identifies the builder as an ecosystem engineer whose niche construction is altering the habitat of every downstream organism in the cognitive ecology.
The Jones-Lawton-Shachak framework distinguishes between autogenic engineers (organisms whose own bodies are the infrastructure—corals, trees) and allogenic engineers (organisms that modify other materials to create infrastructure—beavers, earthworms). The distinction matters because it determines what happens when the engineer is removed. The reef persists when the coral dies; the dam fails when the beaver leaves. Applied to AI: the cognitive infrastructure being built with AI tools is allogenic—it requires continuous maintenance by the engineers that built it. Remove the maintenance and the ecology contracts.
Haeckel's ecology implied engineering without naming it. The relational framework—organism and environment as a single system, neither intelligible without the other—carries the structural implication that every organism's existence alters the conditions for every other. What the Jones-Lawton-Shachak framework added was the category that made the implication operational: engineering as a distinct class of ecological interaction, measurable, comparable, applicable across scales.
The intelligence ecology is being engineered at speed. Every builder using AI tools is modifying the cognitive environment—creating new affordances, eliminating old constraints, altering what downstream users can attempt and what they can think. The Trivandrum engineers in Segal's account did not just change their own work. They changed the habitat their colleagues would inherit, the expectations their managers would form, the conditions under which the next generation of engineers would be trained. Allogenic engineering at organizational scale.
The framework carries a predictive implication. Allogenic engineering creates habitat that depends on continuous maintenance. When the maintenance stops, the habitat reverts—often faster than it formed. The pond behind the dam drains in weeks; the dam took years to build. The cognitive infrastructure being built with AI tools has the same asymmetric durability: fast to construct, fragile without upkeep. The question is what maintains it, and who bears the cost of maintenance.
Clive Jones, John H. Lawton, and Moshe Shachak published 'Organisms as Ecosystem Engineers' in Oikos in 1994. The paper formalized an ecological category that had been implicit in ecological thinking since Haeckel but had never been precisely named or systematized. The framework became standard in ecology within a decade and has since been extended to cultural evolution, technology studies, and—through the Jones-Haeckel lineage this book draws on—AI ecology.
Engineering is a distinct ecological interaction. Organisms that modify habitats do something categorically different from organisms that merely compete for resources.
Autogenic vs. allogenic determines resilience. The distinction between organisms whose bodies are the infrastructure and organisms that build infrastructure externally predicts what happens when the engineer is removed.
Every builder is an engineer. The AI builder's work is ecosystem engineering at organizational scale—altering the cognitive habitat for every downstream organism.
Engineered habitats require maintenance. Allogenic engineering creates infrastructure whose persistence depends on continuous upkeep. When the upkeep stops, the habitat reverts.