Exploring Perspectives of Software Developers on AI Tool Use and Productivity: A Literature Review and Empirical Findings
Year: 2026
Type: article
Abstract: As AI tools shape software development practices and adoption of these tools continues to grow in fast phase, there is an increasing need to better understand software developer productivity from a holistic point of view, beyond narrow and isolated metrics. Hence, this study explores how software developers describe the relationship between the use of AI tools and productivity, focusing on how AI tools are used by software developers and how productivity is experienced across the SPACE framework dimensions (Satisfaction and Well-being, Performance, Activity, Communication and Collaboration, and Efficiency and Flow) to better understand how productivity is experienced in the emergence of AI tools by software developers. The study combines a systematic literature review of 16 recent peer-reviewed empirical studies with semi-structured interviews conducted with four software developers working in companies in Finland. According to the results from these methods, software developers describe AI tools as iterative, interactive and collaborative assistants rather than autonomous productivity enhancers. Productivity gains are mainly perceived in early-stage and knowledge-oriented tasks, where AI tools support exploration, reduce uncertainty and increase perceived efficiency and independence. However, these gains are accompanied by increased monitoring, validation and cognitive effort. Trade-offs emerge across the SPACE dimensions, particularly between short-term efficiency and long-term codebase quality, skills development and well-being. Traditional and narrow activity-based metrics were described as insufficient for capturing redistributed cognitive and supervisory effort. This study concludes that the relationship between the use of AI tools and productivity is best understood as an evolving negotiation between short-term gains and long-term sustainability, shaped by how software developers manage interaction, risk, and quality over time, and by how organisations align expectations, evaluation practices, and support structures with the redistributed nature of work.
Source: Työväentutkimus Vuosikirja
Author: Salojoki, Riikka
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7198453467