Artificial Intelligence (AI) has played a major role in the development and renaissance of sectors in recent years. It has observed the use of AI technologies and tools in its various fields such as medicine, engineering and software development. In the field of software development, AI and its tools have introduced new ways in the development of the tools to help write the code and the possibility of generating it through text commands and providing maintenance and improvement solutions to the code. Our study looks at the impact of code from AI tools using the ChatGPT tool to generate code and does not focus on efficiency as a software quality factor Comparing the code generated by the AI tool to the open source code from the GitHub platform for three tasks of varying difficulty (easy, medium and difficult) Evaluating key quality measures such as annular contractor, number of code lines, duplicate lines, technical debt and code smell/severity results showed that both the tool and the GitHub platform provided effective and efficient code but ChatGPT showed a greater focus on code efficiency, especially in difficult tasks It achieved better results when evaluating the measures of Cyclomatic Complexity, line of code, duplicated lines and technical debt.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Investigating AI-Generated Code on the Impact on Software Efficiency Code Quality Factor

  • Hamed Al-Fawareh,
  • Hazim Mohammad Al-Shdaifat

摘要

Artificial Intelligence (AI) has played a major role in the development and renaissance of sectors in recent years. It has observed the use of AI technologies and tools in its various fields such as medicine, engineering and software development. In the field of software development, AI and its tools have introduced new ways in the development of the tools to help write the code and the possibility of generating it through text commands and providing maintenance and improvement solutions to the code. Our study looks at the impact of code from AI tools using the ChatGPT tool to generate code and does not focus on efficiency as a software quality factor Comparing the code generated by the AI tool to the open source code from the GitHub platform for three tasks of varying difficulty (easy, medium and difficult) Evaluating key quality measures such as annular contractor, number of code lines, duplicate lines, technical debt and code smell/severity results showed that both the tool and the GitHub platform provided effective and efficient code but ChatGPT showed a greater focus on code efficiency, especially in difficult tasks It achieved better results when evaluating the measures of Cyclomatic Complexity, line of code, duplicated lines and technical debt.