Paper Number
ECIS2026-1988
Paper Type
SP
Abstract
Software development is undergoing a paradigm shift driven by Generative AI (GenAI), promised to revolutionize developer productivity. However, this rapid transformation risks exacerbating the long-standing challenge of technical debt. While current literature speculates on these risks, there is a critical absence of empirical, longitudinal evidence quantifying how widespread GenAI adoption is altering standard indicators of code quality. We address this gap through a longitudinal Interrupted Time Series analysis of 1,091 open-source Python repositories. Our analysis reveals a complex, non-uniform impact: post-intervention, small and medium projects saw a statistically significant acceleration in code debt. Conversely, large projects demonstrated a striking trade-off. Their code debt remained stable, while their architectural debt decreased at a significantly faster rate. At the same time, these projects experienced a significant increase in design debt. These findings suggest GenAI’s impact is highly context-dependent, requiring tailored governance strategies for different project scales.
Recommended Citation
Niemeyer, Jonas and Wessel, Michael, "GenAI’s Hidden Cost: An Empirical Study Of AI‑Induced Technical Debt" (2026). ECIS 2026 Proceedings. 7.
https://aisel.aisnet.org/ecis2026/isd_pm/isd_pm/7
GenAI’s Hidden Cost: An Empirical Study Of AI‑Induced Technical Debt
Software development is undergoing a paradigm shift driven by Generative AI (GenAI), promised to revolutionize developer productivity. However, this rapid transformation risks exacerbating the long-standing challenge of technical debt. While current literature speculates on these risks, there is a critical absence of empirical, longitudinal evidence quantifying how widespread GenAI adoption is altering standard indicators of code quality. We address this gap through a longitudinal Interrupted Time Series analysis of 1,091 open-source Python repositories. Our analysis reveals a complex, non-uniform impact: post-intervention, small and medium projects saw a statistically significant acceleration in code debt. Conversely, large projects demonstrated a striking trade-off. Their code debt remained stable, while their architectural debt decreased at a significantly faster rate. At the same time, these projects experienced a significant increase in design debt. These findings suggest GenAI’s impact is highly context-dependent, requiring tailored governance strategies for different project scales.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.