TY - GEN
T1 - Empirical evaluation of automated code generation for mobile applications by AI tools
AU - Aillon, Santiago
AU - Garcia, Alejandro
AU - Velandia, Nicolas
AU - Zarate, Daniel
AU - Wightman, Pedro
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - The rapid advancement of AI technologies has opened up new possibilities for automating various aspects of software development. Mobile app development, in particular, can benefit from AI-powered tools that assist developers in writing code more efficiently, providing suggestions, and reducing the time required for implementation. This document aims to explore how well a modern artificial intelligence tool can assist a mobile application development process. For this work, ChatGPT 3.5 was used to generate a mobile application from scratch using the Flutter framework, while the complete process was evaluated at each step. The evaluation criteria for the experience considered four indicators: code quality, solution quality, response time, and comparison with human-generated code. Results show that, up to a certain level of complexity and by using an interactive process of increasingly detailed prompts, the AI tool is capable of generating functional code, that can be the base for the inclusion of a more complex logic or structure.
AB - The rapid advancement of AI technologies has opened up new possibilities for automating various aspects of software development. Mobile app development, in particular, can benefit from AI-powered tools that assist developers in writing code more efficiently, providing suggestions, and reducing the time required for implementation. This document aims to explore how well a modern artificial intelligence tool can assist a mobile application development process. For this work, ChatGPT 3.5 was used to generate a mobile application from scratch using the Flutter framework, while the complete process was evaluated at each step. The evaluation criteria for the experience considered four indicators: code quality, solution quality, response time, and comparison with human-generated code. Results show that, up to a certain level of complexity and by using an interactive process of increasingly detailed prompts, the AI tool is capable of generating functional code, that can be the base for the inclusion of a more complex logic or structure.
UR - https://www.scopus.com/pages/publications/85186761633
UR - https://www.scopus.com/inward/citedby.url?scp=85186761633&partnerID=8YFLogxK
U2 - 10.1109/C358072.2023.10436306
DO - 10.1109/C358072.2023.10436306
M3 - Conference contribution
AN - SCOPUS:85186761633
T3 - 1st IEEE Colombian Caribbean Conference, C3 2023
BT - 1st IEEE Colombian Caribbean Conference, C3 2023
A2 - Mendoza, Paul Sanmartin
A2 - Navarro, Andres
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 1st IEEE Colombian Caribbean Conference, C3 2023
Y2 - 22 November 2023 through 25 November 2023
ER -