Type: url
Original: https://arxiv.org/pdf/2503.06195
Status: generated
Summary
A PRISMA-based systematic literature review of 90 studies on Human-AI Experience inside Integrated Development Environments (in-IDE HAX), covering work first publicly available between January 2022 and November 2024. It organises findings into three dimensions — Impact (74 of 90 studies), Design (28 of 90) and Quality (19 of 90) — reporting productivity gains alongside verification overhead and over-reliance, three recurring design principles (context awareness, explainability and transparency, user control and adaptability), and correctness, maintainability and security risks in AI-generated code. It also analyses 250 future-work statements and finds the field concentrated on GitHub Copilot and on the implementation stage of the software development lifecycle.