This paper addresses some critical issues still unresolved in AI. Among them, what is the essence of intelligence, what is knowledge, what is understanding and meaning, and how they can be computed, how to build AI that is capable of performing at a human level? To study these topics, the paper introduces a research and development framework called knowledge computing AI (KCAI). The core concept of KCAI is the definition of human-level intelligence as the ability to acquire and apply knowledge and skills. This is a working definition that contains the purpose and plan for studying artificial intelligence. These studies should be explored with a high level of vision. What is this level exactly? As modern industries have proven, mathematical foundations help better understand the underlying phenomena and provide effective solutions. Developing a framework that looks at intelligence from computational knowledge science perspective that examines the computational foundations of knowledge and on its basis develops computational models for acquiring and applying knowledge is critical to advancing human-level AI. This paper is a step in this direction.