Abstract:
Use of AI based tools have gained extensive acceptance by the software development community in the latest past. Various tools have now become an integral part of IDEs. At the same time AI based software development is also emerging with the advancement in generative AI. EngineerGPT, GitHub Copilot and ChatDev are a few examples of such applications. However, effectiveness of these tools still needs to be evaluated throughout the Software Development Life cycle phases helping in developing end-to-end applications. To make comparison about human developer efficiency and AI tools, criteria of feature completeness, quality of code and test comprehensiveness was developed. Two software applications having standard software specifications were given to human developers as well as AI based tool (EngineerGPT). The output generated by both was later compared based on the abovementioned criteria. For Code quality we have used the indicators of Cyclomatic Complexity, Lines of Code (LOC), and Code Duplication. The results of the output reveal that while AIdriven tools efficiently implement core functionality with compact codebases, low duplication, they demonstrate limitations in handling complex requirements and modular design, which can affect code adaptability and feature alignment. In contrast, human developers produce more verbose and modular code, utilizing frameworks and libraries to enhance maintainability. Code based of human developers was large with duplicated code. Test case analysis further highlights differences in coverage, with human-driven approaches achieving complete validation across all requirements, while AI-driven tools effectively cover primary functions but lack thoroughness in secondary features. This research demonstrates the strengths and limitations of AI in software engineering, indicating that while AI-driven tools hold potential for rapid, core functionality development, human expertise remains critical for achieving robust, maintainable, and fully compliant software solutions. The insights contribute to a deeper understanding of AI’s role in software engineering, with implications for optimizing human-AI collaboration in future development workflows.
