The notion of Human Centred Artificial Intelligence systems is receiving much interest in the AI research community. Our work takes an augmented AI perspective, where the primary design goal is to amplify human cognitive power rather than to replace it. In augmented AI the human user works interactively and cooperatively with the AI system. The system learns and adapts to the human user and in turn, the user learns and adapts to the system. Diverse AI techniques, such as machine learning, probabilistic graphical models, ontologies and agent based computing, can be combined to create hybrid AI systems that can adapt to dynamic physical and social environments. In this talk, I will describe some of our previous and current work on AI systems across diverse application domains including health, energy, biodiversity, and earth observation. I will also describe some research challenges and current issues including continual learning, interactive decision making and scientific knowledge discovery, and provides some insights on where we think research on AI systems is heading.