Paper Type
ERF
Abstract
Emerging literature on Generative AI (GenAI) enabled software development outlines a paradox: although GenAI-driven automation speeds up coding tasks, its adoption inadvertently worsens project-level outcomes such as software throughput and stability. We take a process lens and examine how GenAI tools interact with agile practices and hypothesize that this paradox is the result of the vicious cycle of developing higher quantities of code that is more complex and less integrable with existing code, requiring more rework and refactoring effort by developers. We develop a system dynamics model to capture the feedback loops and causal mechanisms underlying GenAI use in agile software projects over time. We plan to collect quantitative project data and qualitative insights from software developers to refine and validate our model. Through this study, we aim to illuminate how GenAI can both elevate individual-level performance and still undercut performance of the agile project, offering theoretical and practical implications.
Paper Number
2274
Recommended Citation
Negi, Kartikeya; Kota, Madhu; Ramesh, Balasubramaniam; and Cao, Lan, "The Paradoxical Impact of Generative AI on Agile Software Development Outcomes: A Process View" (2025). AMCIS 2025 Proceedings. 9.
https://aisel.aisnet.org/amcis2025/it_pm/it_pm/9
The Paradoxical Impact of Generative AI on Agile Software Development Outcomes: A Process View
Emerging literature on Generative AI (GenAI) enabled software development outlines a paradox: although GenAI-driven automation speeds up coding tasks, its adoption inadvertently worsens project-level outcomes such as software throughput and stability. We take a process lens and examine how GenAI tools interact with agile practices and hypothesize that this paradox is the result of the vicious cycle of developing higher quantities of code that is more complex and less integrable with existing code, requiring more rework and refactoring effort by developers. We develop a system dynamics model to capture the feedback loops and causal mechanisms underlying GenAI use in agile software projects over time. We plan to collect quantitative project data and qualitative insights from software developers to refine and validate our model. Through this study, we aim to illuminate how GenAI can both elevate individual-level performance and still undercut performance of the agile project, offering theoretical and practical implications.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
SIGITPROJMGMT