A Framework for the Human in Human-AI Interaction

How can we better account for the ways in which sociocultural structures influence human-AI interactions? In this talk, we discuss a multi-level framework for thinking through the impact of genres of the human on human-AI interaction. We also discuss how we have operationalized this framework using perspectives from the computational and decision sciences, and Black studies with examples from ongoing projects in disaster recovery and Black digital archives. We conclude with reflections on considerations of the human in the engineering lifecycle of AI systems and relations to the multi-level human-AI framework.