AI in Higher Education: Role Domains, Archetypes, and Allocation

Generative AI reduces the time and effort required for many knowledge tasks in universities. AI in higher education can therefore be best understood as an allocation shock, as it can create cognitive and temporal surplus1 and a potential to introduce a sudden increase in usable cognitive capacity. The resulting surplus time and cognitive bandwidth may be allocated somewhere else, for example, teaching, administration, publication throughput, programme-building, or boundary exploration. In this context, the observable “impact” of AI is largely the pattern of reallocation rather than the tool itself. What becomes visible after adoption depends on (1) the role you occupy, (2) what that role is evaluated on, and (3) where the surplus is directed. That is why identical tools produce different institutional outcomes: AI simply amplifies the optimisation function already operating inside the institution. To make this analysable, we describe with role domains, then specify archetypes within those domains. Archetypes in this context are not fixed identities; they are modes that people can combine and may oscillate between.