Content area
AI code-generation tools promise to improve developer productivity, but realizing these gains depends on understanding how developer attributes and work environments interact with these technologies. This quantitative study analyzed professional developers from the 2023 and 2024 Stack Overflow Developer Surveys, conducting confirmatory, exploratory, and predictive analyses to assess the impact of AI code generation on developer productivity, measured as time spent searching for programming solutions. Confirmatory analyses found that AI code-generation usage alone did not significantly reduce search time. However, developer experience, country population, and specific tool–language combinations significantly moderated outcomes. Less experienced developers and developers from smaller-population countries experienced greater efficiency gains. Predictive analyses identified years of professional experience, frequency of workplace interruptions, and country population as the strongest predictors of search behavior. Interaction effects revealed that AI tools such as Codeium and GitHub Copilot influenced productivity differently across programming language environments. Notably, combinations such as Codeium with Systems languages and GitHub Copilot with Rust/R and Ruby were associated with significant changes in search time.
These findings underscore the complexity of AI adoption in professional software development, emphasizing that the benefits of AI code generation depend not only on tool selection but also on developer demographics, experience levels, and technical ecosystems.
