The Configurations of Human Expertise with AI across SoftwareDevelopment Lifecycle

Sammanfattning

Artificial Intelligence (AI) tools are increasingly integrated into softwaredevelopment, transforming how tasks are performed across the SoftwareDevelopment Life Cycle (SDLC). This thesis investigates how human expertise isconfigured alongside AI to support and improve software development processes.Rather than focusing on AI capabilities alone, the study emphasizes the evolving andessential role of human contribution in human-AI systems.The research adopts a qualitative approach and draws on semi-structured interviewswith professionals engaged in software development. A thematic analysis wasconducted using the Human-in-the-Loop (HITL) as an analytical framework, with theSDLC serving as a structural framework to examine how human roles manifest andadapt throughout different development stages. Findings illustrate that human-AIconfigurations are not static or confined to specific stages of the SDLC. Instead,human expertise remains embedded and dynamically distributed throughout thedevelopment process, providing judgment, context awareness, and ethical oversightthat AI alone cannot offer. This structured embeddedness challenges earlier viewsthat human roles diminish as AI advances. This study demonstrates that the relevanceof human expertise is not diminished by AI; rather, it is reinforced by it. Participantsemphasized that HITL is not merely a technical supplement, but a foundationalcondition for reliable, responsible, and contextually meaningful softwaredevelopment.This thesis contributes to a deeper understanding of human-AI collaboration insoftware development by highlighting how the role of human capabilities is not onlypreserved but expanded through interaction with AI tools. It provides actionableinsights for designing development environments where human and AI contributionsare strategically configured for long-term effectiveness.