MINT Seminar 12 Dec: Jake Iain Stone Corporate-Open-Source Relationships in AI

Abstract: The development of open source AI exists at the intersection of market and gift economies, raising fundamental questions about how we should understand and govern these relationships. While scholarship has highlighted problematic dynamics of corporate involvement in open source, focusing solely on these issues overlooks crucial complementarities. I argue that corporate-open source relationships in foundation model development represent a form of mutualism rather than simple co-option. Drawing on Benkler's theory of peer production, I show how foundation models violate key conditions for pure peer production, particularly in their capital requirements and resource allocation challenges. These limitations create an opportunity for productive corporate-community partnerships where companies provide essential infrastructure while benefiting from distributed innovation. However, this mutualism is fragile and can break down through various forms of appropriation that alienate community participants. Understanding these dynamics helps us move beyond seeing corporate involvement as inherently antagonistic toward developing governance frameworks that sustain productive partnerships while guarding against harmful appropriation. This analysis offers practical insights for managing corporate-open source relationships in AI development while contributing to broader debates about the political economy of open source.

Bio: Jake is a PhD candidate in Philosophy with the Australian National University and holds an MSc in data science, a BA (Hons) in philosophy, and a LLB (Hons). Jake's research explores the political economy of artificial intelligence, with particular focus on how compute infrastructure and open source development shape AI innovation and power dynamics. His current work examines institutional arrangements around AI development, including corporate-community partnerships, resource allocation, and governance frameworks.