# An Empirical Study on Usage and Perceptions of LLMs in a Software Engineering Project

**Authors:** Sanka Rasnayaka, Guanlin Wang, Ridwan Shariffdeen, Ganesh Neelakanta Iyer  
**arXiv:** [2401.16186](https://arxiv.org/abs/2401.16186)  
**Date:** 2024-01-29  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2401.16186v1/

## Key points

- Study on use of Large Language Models (LLMs) in academic software engineering projects
- Involvement of 214 participants working in teams of up to six members
- LLMs can enhance productivity in early stages of development
- Adoption varied based on students' coding skills and prior experience with AI generators
- Limitations include incomplete data on AI-generated code annotations and prompts used for generation
- Provides valuable insights into the role of LLMs in software engineering education
- Highlights the need for effective usage of AI tools in teaching programs

## Summary

In their study accepted for publication at the LLM4Code workshop in ICSE 2024, Sanka Rasnayaka, Guanlin Wang, Ridwan Shariffdeen, and Ganesh Neelakanta Iyer delve into the use of Large Language Models (LLMs) in academic software engineering projects. The controlled study involved 214 participants working in teams of up to six members and aimed to explore the impact of LLMs on software engineering projects. Results showed that LLMs can significantly enhance productivity in early stages of development but adoption varied based on students' coding skills and prior experience with AI generators. The study also had limitations such as incomplete data on AI-generated code annotations and prompts used for generation. Despite this, it provides valuable insights into the role of LLMs in software engineering education and highlights the need for effective usage of AI tools in teaching programs. have potential to improve  by enhancing , but proper utilization is crucial for success.