From Commons To Code Generation

How AI-Assisted Development is Eroding the Open Source Ecosystem

Abstract

This dissertation examines the structural relationship between the rise of artificial intelligence coding tools and the long-term sustainability of open source software (OSS) as a sociotechnical institution. Drawing on the theoretical resources of Science and Technology Studies — including Benkler's commons-based peer production, Ostrom's commons governance framework, and STS concepts of interpretive flexibility, technological momentum, and the politics of artifacts — it argues that AI-assisted development is systematically eroding three conditions that have historically sustained OSS: the behavioral practice of code reuse and contribution, the informal pipeline through which contributors are formed, and the existence of shared problems that make collective solutions worthwhile. The dissertation engages seriously with counterarguments, examines the implications of OSS decline for the software industry, innovation ecosystems, and governance, and argues that deliberate institutional and policy intervention is required to prevent the irreversible hollowing out of the digital commons.