In psychology, a cognitive map is the mental model we build of the world around us, how things connect, where we are in relation to other points, what routes are possible when conditions change.
Edward Tolman coined the term in 1948, showing that rats in a maze do not just follow a memorized path; they build an internal map that lets them take shortcuts when something is blocked. Humans do this too, not only in physical space but in conceptual space: we build maps of ideas, systems, institutions and relationships.
Tools like GPS, search engines and now #AI are powerful partly because they let us offload some of that mapping work. But research on navigation shows a trade off: when we rely too much on turn by turn directions, our own spatial mapping skills weaken.
We get to the destination, but we do not really know where we are.
The same risk exists with AI for thinking and learning: if we always let a model decide what is relevant, how to connect ideas or which next step to take, we risk losing our own sense of the terrain.
The point is not to preserve every map. It is to preserve the ability to make one.
#learning
