Abstract

AI-based software presents unique challenges—stochastic behavior, opaqueness, continuous learning—those traditional requirements engineering (RE) struggles to accommodate. In this paper, we propose an AI-driven Requirements Engineering Framework (AI-RE) tailored to address these challenges across the RE lifecycle. Grounded in ISO/IEC 25010 and SC-42 standards, AI-RE integrates modular layers for human needs, model behavior, data quality, explainability, validation metrics, and governance. We evaluate AI-RE in two domains—autonomous pedestrian detection and VR video enhancement—demonstrating improvements in requirement completeness (92% vs 70%), explainability ratings (4.2 vs 2.8 Likert), and stakeholder satisfaction (4.5 vs 3.5). Comparative analysis with RE4HCAI, GenAI-NFR, and KAOS shows AI-RE’s distinct strengths in traceability, iterative validation, performance awareness, and explainability. Limitations include complexity and tool support needs. Future directions include toolchain integration and large-scale evaluations.

References

1. Sutcliffe, A. G., & Sawyer, P. (2013). Requirements elicitation: Towards the unknown unknowns. Requirements Engineering, 18(2), 138–153. https://doi.org/10.1007/s00766-012-0161-8
2. Breaux, T. D., & Antón, A. I. (2008). Analyzing regulatory rules for privacy and security requirements. IEEE Transactions on Software Engineering, 34(1), 5–20. https://doi.org/10.1109/TSE.2007.70766
3. Doshi-Velez, F., & Kim, B. (2017). Towards a rigorous science of interpretable machine learning. arXiv preprint arXiv:1702.08608. https://arxiv.org/abs/1702.08608
4. Broy, M. (2006). Challenges in automotive software engineering. In Proceedings of the 28th International Conference on Software Engineering (pp. 33–42). ACM. https://doi.org/10.1145/1134285.1134293
5. ISO/IEC 9126-1:2001. (2001). Software engineering — Product quality — Part 1: Quality model. International Organization for Standardization.
6. IEEE Global Initiative. (2019). Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems (1st ed.). IEEE Standards Association.

Download PDF
Back