1 item with this tag.
The predominant AI interface paradigm — text boxes and chronological chat histories — reproduces a deeply embedded cognitive metaphor that misaligns with how professional expertise develops. Drawing on Lakoff and Johnson's container schema, this essay traces how a single organising metaphor has been uncritically reproduced across physical, digital, and AI-mediated learning environments, artificially enclosing knowledge that practitioners must mentally reintegrate. Rather than proposing to replace bounded learning spaces, this essay explores graph-based learning environments as an alternative paradigm where bounded spaces become visible communities within a navigable network. AI serves as both conversational partner and network weaver, with conversations spatially anchored to relevant concepts rather than isolated in chronological chat histories. This reconceptualisation — from enclosure to constrained traversal — suggests possibilities for AI-supported learning environments that better develop the integrative capabilities defining professional expertise.