Beginning in the late 1980s, computer security researchers began discussing the risks associated with ``software monocultures.'' Within a decade, this metaphor had gained such prevalence that it could be invoked as self-evident, taking for granted that the industry should ``avoid monoculture in computer operating systems'' for reasons ``just as reasonable and obvious as avoiding monoculture in farming.'' This paper explores how the agricultural metaphor of ``monoculture'' migrated into computing discourse, naturalizing discussions of technical vulnerabilities and centralization patterns. Building on research from science and technology studies (STS), the authors argue that the monoculture metaphor is epistemologically significant, both describing and shaping computing practices. Rather than accepting the simplified narrative that monocultures represent only technical risks, the authors draw on agricultural history to develop a more nuanced understanding of monocultures as deeply entrenched systems of power relationships characterized by dependency, monopoly control, and systemic lock-in. The authors extend this analysis to the current development of a new monoculture of computing forming with the development of energy and resource intensive AI systems and infrastructure.
ACM CHI Conference on Human Factors in Computing Systems