Human-machine design considerations in advanced machine-learning systems

This paper explores issues related to human-computer interaction with the new class of machine-learning systems that represent an exciting development on the frontiers of information technology. These systems represent a significant breakthrough in humanity's decades-long endeavor to build computers that are "more like us''-fundamentally designed to map to and support human abilities and where the needs of people are central to the design process. This paper examines the human aspects of developing such new technology, with a focus on question-and-answer machine-learning systems. Considerations such as how human behavior should be addressed in the design and development of such systems are presented, followed by a series of potential application domains for the new technology.