Towards the Development of Artificial Intelligence-based Systems: Human-Centered Functional Requirements and Open Problems

Abstract

The increasing capability of AI-powered systems, including self-aware and unmanned systems, at automating simple to sophisticated tasks and their wide areas of real-world interventions in boosting productivity and enhancing competitiveness has offered transformative potentials leading to better quality of life. These systems have lately become an inseparable part of human lives. However, incorrect use leading to unintended consequences, safety, fairness, trustworthiness are major concerns of these emerging ubiquitous systems. In this paper, an attempt was made to concisely present and discuss key human-centered functional requirement specifications of emerging Artificial intelligence-based systems especially interpretability, explainability, fairness, transparency and security. Some emerging toolkits and open libraries for developing and evaluating AI-based systems are also discussed. A number of open problems with respect to managing the tradeoff among these requirements and systems’ performance are presented to guide future researches in this direction.