AI and the Restructuring of the Software Development Market: Evidence from Open Source Ecosystems

Open source software development, a pillar of collective innovation and production of digital commons (Lerner & Tirole, 2000; Lakhani & von Hippel, 2003), is undergoing an unprecedented transformation driven by generative artificial intelligence (genAI). This is profoundly reshaping code production practices and contributors’ incentives. In parallel, voluntary contributions have gradually declined, while the participation of major digital players has intensified, reflected in the rise of salaried contributors, “open core” models, and service-oriented activities (Li et al., 2024). While the participation of large corporations provides resources, professional workflows, and visibility to the open source community, it is also perceived as a potentially undermining factor from the perspective of independent contributors, whose incentives are often driven by non-monetary considerations such as altruism, learning, recognition, enjoyment, or a sense of community belonging, raising questions about governance, technological orientation, and evolving collaboration patterns (Bitzer et al., 2007; Meissonier et al., 2010; Zhang et al., 2024). The purpose of this thesis is to examine how the advent of genAI is affecting open source software development more broadly, with particular emphasis on its potential to recalibrate or disrupt the delicate dynamics between independent and corporate participation in open source. Existing work shows that AI-assisted coding can increase productivity and code quality while recomposing tasks (Hoffmann et al., 2024; Yeverechyahu et al., 2024; Cui et al., 2025; Peng et al., 2023), and that online communities exposed to genAI show changes in participation, specialization, and quality (Burtch et al., 2024; Quinn & Gutt, 2025). Building on a large-scale GitHub data collection on repositories, contributors, and affiliations, the first chapter of the thesis measures whether paid and unpaid developers reacted differently to the release of ChatGPT-3.5. We analyze how genAI interacts with intrinsic motivations and extrinsic rewards, and whether assistance tools have a different impact on professional routines compared to voluntary exploration. The next stages will be to study how genAI reshapes the nature of contributions, the orientation of projects, and the robustness of the resulting code, in order to identify which groups capture AI-enabled productivity gains and the implications for the open source ecosystem. Empirically, we rely on three complementary data sources. A panel of developer-month observations summarizing the number of contributions by type (pushes, pull requests, issues, etc.). Then is a corpus of repository-descriptions, which allows us to conduct a topic analysis to assess changes in the functionalities of open source projects. Finally, we will use a database linking contributors and the code they produce, allowing us to assess the impact of genAI on code-quality and related measures documented in the software engineering literature (Yeti¸stiren et al., 2023).