[Context] Present-day IT systems are more and more dependent on artificial intelligence (AI) solutions. Developing AI-based systems means facing new challenges, not known for more conventional systems. Such challenges need to be identified and addressed by properly adapting the existing development and management processes. [Objective] In this paper, we focus on the requirements engineering (RE) area of IT projects and aim to propose the RE process that would be able to address at least some of the reported challenges. No proposal of such process could be found in the existing literature. [Method] We conducted a literature review using a snowballing technique to identify RE-related challenges for AI-based systems. Then, we compared several RE industry guides, selected a well-established RE process and adapted it by introducing additional practices. The additional practices were proposed as result of brainstorming and ideation process. [Results] The contributions of this paper include: a list of identified challenges, a set of additional practices to mitigate challenges and a model of the adapted RE process which integrates such practices. [Conclusions] The proposed process is available for validation activities and can be used by researchers and practitioners as a base for further adaptations of RE approaches to AI solutions.
Autorzy
Informacje dodatkowe
- Kategoria
- Aktywność konferencyjna
- Typ
- publikacja w wydawnictwie zbiorowym recenzowanym (także w materiałach konferencyjnych)
- Język
- angielski
- Rok wydania
- 2023