Developer Productivity with GenAI

TL;DR

This paper investigates how GenAI adoption affects developer productivity across multiple dimensions by applying the $SPACE$ framework to a large-scale survey of 415 practitioners. Using a survey instrument adapted to GenAI contexts and analyzing responses by AI usage frequency, the study reveals that overall perceived productivity changes are limited, with frequent users reporting modest gains in Efficiency and Flow and in Satisfaction, but no consistent improvements in Performance, Activity, or Communication. The findings challenge the assumption that GenAI accelerates software development across the board and highlight potential spurious productivity, emphasizing the need for nuanced, multi-dimensional measurement and governance. Practically, the work suggests integrating GenAI with clear goals and collaborative processes, and points to directions for extending the SPACE framework to better capture AI-mediated work dynamics.

Abstract

Generative AI (GenAI) tools are increasingly being adopted in software development as productivity aids. However, evidence regarding where and when these tools actually enhance productivity is unclear. In this paper, we investigate how GenAI adoption affects different dimensions of developer productivity. We surveyed 415 software practitioners to capture their perceptions of productivity changes associated with AI-assisted development using the SPACE framework - Satisfaction and well-being, Performance, Activity, Communication and collaboration, and Efficiency and flow. Our results, disaggregated by frequency of AI usage, reveal limited overall productivity change, highlighting the productivity paradox in which developers become faster but do not necessarily create better software or feel more fulfilled.

Figures (6)

• Figure 1: Split-violin plots of aggregated SPACE scores. Left half = non-frequent AI users (blue), right half = frequent AI users (red). Boxplots show median (line) and mean (dot). Gray band marks the neutral ("no-change") range with green for positive and red negative perceptions; example tick labels shown for Satisfaction & Well-being (4 items: 12 = all Neutral, 16 = all Agree).
• Figure 2: Distribution of responses to Satisfaction and well-being dimension items.
• Figure 3: Distribution of responses to Performance dimension items.
• Figure 4: Distribution of responses to Activity dimension items.
• Figure 5: Distribution of responses to Communication and collaboration dimension items.