From Typist to Collaborator: A Systematic Review of AI’s Impact on Programming Efficiency and the Shift in Developer Roles
Main Article Content
Keywords
artificial intelligence, programmer, project development
Abstract
With continuous advancements in AI and large language models, AI in programming has evolved from simple code completion to intelligent agents with contextual understanding. This change has not only had a significant impact on the field of programming but also on programmers' employment. This paper aims to explore new working models for future programmers and analyze the efficiency and potential risks of AI programming. This paper systematically reviews key empirical studies from relevant papers published between 2020 and 2025 and finds that AI exhibits significant efficiency issues in programming. On the one hand, in LeetCode, AI performs well on low-to-medium difficulty problems, and AI programming has already surpassed 85% of human programmers in algorithm competitions. On the other hand, the actual development efficiency of AI when handling large open-source projects actually decreases by approximately 19%, and a series of safety and efficiency issues are found in AI programming, such as code redundancy and over-coupling when writing large projects. This article explores the transformation of programmers from code writers to AI collaborators, examining the advantages and disadvantages of AI in programming. It also elaborates on the specific requirements for future programmers, such as a deeper understanding of the nature of tasks and the need to maintain a dynamic balance between programmers and AI to achieve greater efficiency in project completion.
References
- [1] Liang, S., Jiang, N., Hu, Y., & Tan, L. (2025). Can Language Models Replace Programmers for Coding?. Association for Computational Linguistics.
- [2] Kuhail, M. A., Mathew, S. S., Khalil, A., Berengueres, J., & Shah, S. J. H. (2024). “Will I be replaced?” Assessing ChatGPT’s effect on software development and programmer perceptions of AI tools. Elsevier.
- [3] Hou, W., & Ji, Z. (2024). Comparing large language models and human programmers for generating programming code. Wiley.
- [4] Wang, H., Gong, J., Zhang, H., Xu, J., & Wang, Z. (2024). AI Agentic Programming: A Survey of Techniques, Challenges, and Opportunities. Association for Computing Machinery.
- [5] Hassan, A. E., Oliva, G. A., Lin, D., Chen, B., & Jiang, Z. M. (2024). Towards AI-Native Software Engineering (SE 3.0): A Vision and a Challenge Roadmap. arXiv.
- [6] Becker, J., Rush, N., Barnes, E., & Rein, D. (2025). Measuring the Impact of Early -2025 AI on Experienced Open-Source Developer Productivity. arXiv.
- [7] Nguyen-Duc, A., Cabrero-Daniel, B., Przybylek, A., Arora, C., Khanna, D., Herda, T., Rafiq, U., Melegati, J., Guerra, E., Kemell, K. -K., Saari, M., Zhang, Z., Le, H., Quan, T., & Abrahamsson, P. (2025). Generative Artificial Intelligence for Software Engineering—A Research Agenda. Wiley.
- [8] Liu, V., & Chilton, L. B. (2022). Design Guidelines for Prompt Engineering Text-to-Image Generative Models. Association for Computing Machinery.
- [9] Pearce, H., Ahmad, B., Tan, B., Dolan-Gavitt, B., & Karri, R. (2025). Asleep at the Keyboard? Assessing the Security of GitHub Copilot’s Code Contributions. Association for Computing Machinery.
- [10] Weisz, J. D., Kumar, S. V., Muller, M., Borromeo, K. E., Goldberg, A., Heintze, K. E., & Bajpa, S. (2025). Examining the Use and Impact of an AI Code Assistant on Developer Productivity and Experience in the Enterprise. Association for Computing Machinery.
- [11] Caballar, R. D. (2020). Programming Without Code: The Rise of No-Code Software Development. Institute of Electrical and Electronics Engineers.
- [12] Digital Life Kazik. (2026, March 3). We have witnessed history once again. WeChat Official Account. https://mp.weixin.qq.com/s/MTnpLe4KxTU0B9PgKa4Yyw
Article Sidebar
This work is licensed under a Creative Commons Attribution 4.0 International License.