How does the use of generative AI tools influence knowledge practices among software engineers? This is the question Victoria and Julie try to answer in the empirical study of programmers who use GenAI.
Julie and Victoria delivered their master’s thesis last week. The thesis is based on 14 semi-structured interviews with software developers with varying experience. Victoria and Julie use Nonaka and Takeuchi’s knowledge transfer model to analyse the role of GenAI in collaborative processes of software engineering.
- Title: An Efficiency Booster or a Problematic Shortcut? An empirical study exploring the impact of generative AI on software engineering knowledge practices.
- Authors: Victoria Eilertsen and Julie Vist
- Abstract: The software engineering profession is currently facing distinct pressures, characterized by widespread layoffs, fewer job opportunities, and declining student enrollment, with Generative AI (GenAI) widely discussed as a contributing factor. Due to its ability to synthesize information, GenAI presents a disruption to software engineering, which is a fundamentally knowledge-intensive profession. This qualitative study investigates how the integration of GenAI into the daily workflows of software engineers influences existing knowledge practices. Using Nonaka and Takeuchi’s (1995) SECI model as an analytical lens, this research examines how GenAI reshapes knowledge creation across the four knowledge conversion modes: socialization, externalization, combination, and internalization. Based on semi-structured interviews with industry professionals, the findings reveal that GenAI fundamentally alters several dimensions of the SECI framework. Notably, GenAI often serves as an automated knowledge source, reducing traditional socialization among team members. Conversely, it enhances externalization by challenging engineers to articulate their tacit knowledge into explicit prompts. Furthermore, GenAI accelerates combination through its ability to structure and synthesize vast amount of data. Finally, while GenAI’s customized outputs can foster deeper internalization, the immediacy with which it generates answers risks positioning engineers in a more passive role, potentially hindering long-term knowledge assimilation. Ultimately, this research offers critical insights for organizational leaders establishing GenAI governance and provides software engineers with a reflective framework to understand how GenAI usage influences their professional development and knowledge creation.