A Review of AI-Assisted Impact Analysis for Software Requirements Change: Challenges and Future Directions
Ahmad Abdelhafiz Ali Samhan, Suhair AlHajHassan, Sara Abu Dabaa’t, Ali M. Elrashidi · 2024
This paper presents a review study that explores the landscape and challenges of AI-assisted impact analysis in software requirements engineering. The growth of complex and change-prone software systems has highlighted the need for AI techniques—especially machine learning (ML) and natural language processing (NLP)—to be harnessed to improve impact analysis scalability and efficiency. Key challenges identified include data quality issues, scalability constraints, and ethical problems related to bias and accountability. To investigate these challenges and provide insights, a review was conducted across IEEE Xplore, ACM Digital Library, and Google Scholar, cross referencing between keywords that include concepts from the domains of "software requirements", "impact analysis", and "AI-Assisted approaches". Inclusion criteria focused on studies that presents empirical data on AI-assisted impact analysis. This study highlights research gaps and presents insights for overcoming obstacles using emerging technologies such as edge AI in order to enhance post-impact analysis. Findings show that while AI improves traditional impact analysis, further work is necessary to deliver effective, ethical and practical implementations within evolving software contexts.