Challenges in Developing Artificial Intelligence-based Systems using Classic Software Development Life Cycle

Abstract

Artificial intelligence (AI) tools have become an essential part of modern software systems, driving automation, predictive analysis, and intelligent decision-making. However, the development of AI tools using traditional software engineering methodologies, specifically the Software Development Life Cycle (SDLC), presents unique challenges. This paper explores the intricacies and limitations of the classic SDLC in the context of AI-based system development, covering issues related to requirements gathering, design, implementation, testing, deployment, and maintenance. The analysis highlights the need for flexibility and iterative development models to meet the dynamic nature of AI projects. This paper summarises possible considerations that need to be taken care of since building AI-based systems requires a more flexible and iterative approach to accommodate the uncertainty, complexity, and dynamism in its own development.