SPACE: Human Agency in the Development of Thought with AI

Generative AI can now participate in activities that are closely connected with thinking. It can formulate an unfinished idea, question an assumption, interpret what someone has written, compare different positions, and suggest what may follow. These capabilities can be useful, but the quality of the resulting output does not necessarily show what has happened to the person’s own participation in the thinking process. This paper introduces SPACE, a conceptual framework for examining human agency as thought develops in interaction with AI. The framework follows three connected processes: articulation, reflection, and transformation, expressed within SPACE as Talk, Think, and Form. Articulation concerns the process of giving form to something that may not yet be clearly understood. Reflection concerns examining what has been articulated, including assumptions, tensions, evidence, recurring questions, and changes over time. Transformation concerns the point at which an emerging understanding begins to affect subsequent choices, actions, creations, inquiries, or experiences. Based on these processes, SPACE proposes three corresponding forms of human authority. During articulation, authorship of emerging meaning concerns the person’s participation in giving an unfinished thought its initial form. During reflection, interpretive authority concerns the person’s role in deciding what observations, patterns, and possible interpretations mean. During transformation, consequential authority concerns the person’s role in determining what significance an emerging understanding will have for what follows. These categories are proposed as analytical constructs rather than empirically established components of cognition. SPACE does not argue that AI should be removed from human thinking. It provides a way to examine how AI can participate in the development of thought while the person continues to take part in forming, interpreting, and carrying that thought forward.