Cybernetics

Cybernetics is the study of how systems pursue goals through feedback, communication, and recursive exchange. Norbert Wiener named the field in 1948, taking the term from the Greek word for a steersman, and it has since been described as a science of purpose: it asks how a system senses the difference between where it is and where it aims to be, corrects its course, and revises its aims in the process. A second wave, associated with Heinz von Foerster and Gordon Pask, placed the observer inside the system being observed and turned attention to conversation as the medium through which participants come to shared understanding. Design cybernetics extends that tradition into the making of interactive systems, with Hugh Dubberly and Paul Pangaro arguing that the task of design is to create conditions in which others can design and act. Cybernetics shares much of its explanatory apparatus with semiotics, and the two traditions met in the twentieth century across biosemiotics, systems theory, and theories of cognition.

Anthropology helped shape this vocabulary at its founding. Margaret Mead and Gregory Bateson took part in the Macy Conferences on Cybernetics, where the concepts of feedback and circular causality were worked out across disciplines, and Bateson carried those ideas into his later work on mind and ecology. The discipline then largely moved past the tradition, even as its core commitments, relational thinking and attention to how parties in an exchange shape one another, remained anthropological through and through. Matt Artz and Lora Koycheva trace this history in the chapter An Anthropological History of AI, and Matt presented it in the webinar Anthropology and AI: From Past Engagements to the Contemporary Landscape.

Cybernetics is the organizing frame for the Human Organization special issue on anthropology and AI that Matt co-edited with Lora Koycheva. Their introduction, Who Designs the Loop? Anthropology and the Cybernetics of Human–AI Relations, describes human-AI relations as a loop, a recursive circuit running from anthropologist to system to institution to consequence and back. It argues that large language models have made cybernetics newly actionable, because such systems can be steered in real time through natural language, and it distinguishes the conversational loop now taking shape from a dialogical loop toward which designers should deliberately steer. Matt’s concluding commentary, Closing the Loop: Relocated Expertise and Building as Anthropological Practice, takes up the practical half of that question: what it would take for anthropologists to design the loop rather than only diagnose it.

The same frame runs through Matt’s other writing on artificial intelligence. The Magical Power of Words in Large Language Models treats AI not as a singular tool but as a mode of forming a mutual cybernetic relationship in which human expertise and computational pattern recognition shape one another. Multi-Agent Ethnography draws on evidence of AI functioning as a cybernetic teammate to argue for purpose-built research agents designed with anthropological considerations. Together with design anthropology, which likewise takes intervention rather than description as its starting point, cybernetics gives Matt’s work its central question: who gets to steer the exchange between humans and AI, under what conditions, through which infrastructures, and toward whose goals.