Co-becoming is a theoretical framework for understanding the mutual transformation that occurs when humans and AI systems work together. Rather than treating human-AI interaction as a one-directional relationship in which humans use tools or tools replace humans, co-becoming describes a process in which both the human and the system are changed through sustained engagement. There is no fixed end state. The process is perpetual, emergent, and shaped by the specific conditions under which human and machine encounter one another.
Intellectual lineage
The concept draws on two established lines of thought. Donna Haraway’s work on companion species describes the mutual becoming of organisms that live in sustained relation to one another, where neither partner remains unchanged by the encounter. Lucy Suchman’s research on human-machine reconfigurations examines how the boundaries between human and technical agency are drawn and redrawn in practice. Co-becoming extends both into the specific domain of AI, where the pace, scale, and autonomy of computational systems create new dynamics that earlier frameworks did not anticipate.
Four characteristics
Co-becoming is defined by four characteristics. Mutual adaptation describes how both human and AI adjust their behavior in response to each other over time. Behavioral interdependence captures the way actions by one party shape the possibilities available to the other. Emergent dynamics refers to outcomes that neither human nor AI would produce alone, arising from the interaction itself. Collective transformation describes the broader shift that occurs when many human-AI pairs co-become simultaneously, reshaping institutions, practices, and expectations.
Design principles
Co-becoming also generates three design principles for building equitable human-AI systems. Transparent evolution holds that the system’s changes should be visible to its users. Intentional agency requires that humans retain meaningful choice in the relationship. Collective benefit asks that the mutual transformation serve the community, not just the platform. Together, these principles distinguish co-becoming from the default trajectory of most AI systems, in which adaptation happens silently and primarily serves commercial objectives.
Applications
The framework has been applied to design research, where co-becoming reconceives design anthropology as a continuous process of mutual becoming rather than a series of punctuated participatory events. It also informs the development of recommender systems that use behavioral capital to reward participation rather than reinforcing existing hierarchies of attention and visibility.
Origin
Matt Artz developed the concept of co-becoming and published it in the Routledge volume Anthropology and AI (2026), in a chapter titled Co-Becoming with AI: An Anthropological Framework for Equitable Human-AI Systems. The framework is a central theoretical contribution of his PhD dissertation at Aalborg University’s Techno-Anthropology Lab and informs his broader work in AI Anthropology.