一句话：博士研究提案，旨在理解并优化生成式 AI 在限时编程挑战中的集成，以平衡协同智能、技能发展与评估有效性。
软件工程正因生成式人工智能工具而经历深刻变革，但这类工具对限时编程挑战（如竞赛编程、黑客松和技术面试）中程序员表现与学习结果的影响仍缺乏系统理解。限时编程挑战既是教育评估也是职业招聘的重要环节，引入 GenAI 后，技能发展、评估效度与编程专长本质都面临关键问题。现有研究缺少对 GenAI 如何影响限时环境下表现与学习成效的系统认识，导致教育机构与行业难以制定使用指南。若缺乏合理的集成框架，可能带来两类风险：一是过度依赖 AI，损害基本技能发展；二是传统评价指标与新兴“协同智能”编程实践错位。该博士研究提案旨在探索上述问题的潜在解决方案。
The software engineering landscape is experiencing a significant transformation with the advent of Generative Artificial Intelligence (GenAI) tools. While these tools promise to enhance programmer productivity, their impact on Time-Constrained Coding Challenges (TCCs), such as competitive programming, hackathons, and technical interviews remains poorly understood. TCCs are fundamental to both educational assessment and professional recruitment in software engineering. Integrating GenAI tools in these environments raises critical questions about skill development, assessment validity, and the nature of programming expertise. Current research lacks a systematic understanding of how GenAI tools influence programmer performance and learning outcomes in time-constrained environments. This gap is particularly concerning as educational institutions and industry organizations struggle to establish guidelines for GenAI use in TCCs. Without proper frameworks for integration, there are risks of: (1) Over-reliance on AI, potentially compromising fundamental skill development, and (2) Misalignment between traditional evaluation metrics and emerging co-intelligent programming practices. This doctoral research proposal aims to explore potential solutions for these issues.
@inproceedings{paperbot8980,
title = {Co-Intelligence in Software Engineering: Understanding and Optimizing GenAI Integration in Software Engineering for Skill Development in Time-Constrained Programming},
author = {Omkar Joshi},
booktitle = {FSE},
year = {2025},
doi = {10.1145/3696630.3731469}
}