Impact of Generative AI Assistants on Software Engineers' Skills, Workflow Efficiency, and Decisions

The growing adoption of generative AI assistants such as GitHub Copilot and ChatGPT is reshaping software engineering practices. While these tools promise faster development and reduced repetitive work, their effects on developer skills, workflow efficiency, and architectural decision-making remain insufficiently examined. This qualitative study explores the interactions between professional software engineers and AI-assisted programming tools, as well as their impact on technical skills, confidence, and system design practices. Data were collected through in-depth interviews with representatives from diverse organizations and observational analysis of AI-assisted coding sessions. Findings reveal a complex interplay of skill augmentation and cognitive offloading, alongside risks of deskilling, shifts in architectural reasoning, and overreliance on AI-generated solutions.