Heterotopic AI Literacy: A Pedagogy of Epistemic and Temporal Disjunction

Ilkka Tuomi's 5-A model of knowledge creation situates cognition within collective, temporally structured processes of sensemaking. Drawing on Bergson, Dewey, Vygotsky, and traditions of autopoiesis and anticipatory systems associated with Maturana, Varela, and Rosen, this account treats knowledge as an emergent property of ongoing activity rather than as a static resource. Processes of appropriation, articulation, accumulation, anticipation, and action describe how meaning is produced and reorganized over time within socio-cultural systems shaped by duration, memory, and collective practice.

A second framework, articulated by N. Katherine Hayles, specifies cognition at the level of the system while also developing a distinct account of temporality. She defines cognition as a process that interprets information in contexts connected to meaning, and proposes criteria—sensing, interpreting, responding flexibly, anticipating, and learning—to distinguish cognition from adaptation. Within this framework, temporality enters both through anticipation and learning, which reorganize a system’s relation to meaning over time, and through a distinction between systems governed by entailing laws and those characterized by enablement, where the space of possible emergence is itself subject to transformation.

These frameworks operate at different levels but their intersection reveals a breakdown in shared temporal coordination, since each isolates distinct epistemic dynamics that unfold according to different temporal regimes. Tuomi’s account describes how meaning emerges, accumulates, and reorganizes across collective systems, while Hayles provides criteria for identifying where interpretive processes occur and how the conditions of possibility shift within those processes.

In AI-mediated spaces, these levels intersect within a shared process without resolving into a single temporal frame. Human cognition remains embedded in historically and culturally structured practices of meaning-making, carrying duration and collective memory across extended temporal horizons, while LLMs operate via constraint fields shaped by prior linguistic traces that generate responses within probabilistic landscapes of continuation. Both participate in anticipatory processes that reorganize informational relations in the exchange, even as their temporal dynamics remain materially distinct.

AI-mediated learning environments are not just distributed systems; they are heterotopically disjunctive sites where multiple temporal logics coexist without a shared metric of coordination, producing structural misalignments in how meaning persists, accumulates, and becomes accountable. A heterotopic AI literacy emerges where knowledge is produced through materially and temporally distinct processes that resist reduction to a single regime of understanding and lack any stable basis for reconciling meaning over time.