The Design of AI Algorithms for DG Setting
N Ganeshkumar, Purushottam Das, Laith H. Jasim Alzubaidi, Dharmesh Dhabliya, Wesam Abed A L Hassan Alhaidry, Muhsin J. Jweeg, Ahmed Ansari, Ali SaadAlwan · 2024
Artificial intelligence (AI) is the bedrock of management systems and will continue to support human society both now and in the future. These systems’ decision-making processes must be transparent, especially in sensitive areas such as autonomous driving and healthcare, to build trust. Although methods exist for achieving algorithmic transparency, there are no well-established design techniques to systematically address descriptive issues when reviewing AI. In response, we introduce the Explainability in Design (EID) approach, which operationalizes the definition of AI into six fundamental concepts. EID was refined through iterative empirical studies, including applied learning and interviews with practitioners. An empirical study with AI system designers tested the utility of EID, demonstrating its effectiveness in reducing barriers to integrating explanatory functionality in AI systems and enhancing decision-making processes.