The perverse outcomes that arise when interventions designed to solve a problem create incentive structures producing more of the problem — the policy analog of Gibson's emergent affordance behavior.
The cobra effect takes its name from an apocryphal colonial-era episode in which British authorities in Delhi, alarmed by cobra populations, offered a bounty for dead snakes. Enterprising citizens began breeding cobras to collect bounties. When the government discovered the breeding and canceled the program, the breeders released their now-worthless snakes, producing a larger cobra population than existed before the intervention. Whether or not the incident happened, it names a real structural phenomenon: interventions interact with the incentive structures of their environment to produce outcomes inverse to those intended. The cane toad disaster is its ecological cousin. AI tool deployment has produced its own cobra effects: tools designed to increase productivity that generate compulsive engagement; tools designed to free users from tedious work that colonize the hours formerly available for rest; tools designed to democratize capability that consolidate power in those who own the infrastructure. The effect is not cynical prediction. It is structural analysis of what happens when designed interventions meet complex ecological systems.