CONCEPT
Open Worlds and Closed Worlds
Suchman's distinction between the bounded, representable domains in which AI systems succeed and the unbounded, emergent reality in which human practitioners must act — the structural boundary where situated intelligence begins.
Open worlds and closed worlds is Suchman's analytical distinction between the bounded domains in which computational systems operate successfully and the unbounded, emergent reality in which human practitioners must actually live and act. A closed world is one in which the variables are known, the contingencies are bounded, and a plan can specify the action in advance: the chessboard, the training corpus, the described situation. An open world is one in which the variables cannot be fully enumerated, the contingencies are unbounded, and action must be improvised in response to what the actor actually encounters: the deployment environment, the courtroom, the battlefield. AI systems — no matter how sophisticated — operate on representations of open worlds, and the boundary between representation and reality is where situated human intelligence has always lived.
In The You On AI Field Guide
The distinction is central to understanding what AI can and cannot do. In her 2025 AI Now Institute interview,