What does AI-assisted development look like in a big open-source project?

A recent analysis by Unleash explores AI-assisted development within their open-source projects, measuring the extent to which artificial intelligence is integrated into coding practices. By labeling pull requests (PRs) with a new AI usage scheme, the team found that 57.2% of the PRs involved some form of AI assistance, highlighting a preference for collaborative engagement with AI tools rather than complete reliance. Categories revealed that refactoring and documentation primarily attracted AI usage, while human developers maintained control over bug fixes and maintenance tasks where context and insight are essential.

The study also identified distinct developer archetypes based on their AI utilization, ranging from "AI First" to "AI Minimal," illustrating significant variability in individual approaches to AI integration. Key findings indicate that AI tools excel with predictable, medium-complexity tasks, while humans generally handle both simple fixes and larger refactors. This comprehensive analysis not only contributes to a clearer understanding of AI’s role in software development but also sets the stage for future investigations into long-term trends and cross-language comparisons in AI usage.