Temporal Aspects of Human-AI Collaborations for Work

In this position paper we highlight the importance of temporal aspects in human-AI collaboration for work. We focus on chat-based interactions powered by Large Language Models (LLMs), because of their increasing availability and extensive use for knowledge-based and creative work. We present four areas in which the different temporal characteristics of humans and of LLMs influence the interaction between them: interaction pace, temporal context, continuity over time, and rhythm and attunement. In each area we discuss these differences and their pitfalls and potentials, point to current choices made by the designers of LLM interfaces, make initial speculations about possible design opportunities, and suggest open research questions in the context of work. Finally, we propose a number of LLM agent personas that integrate different temporal characteristics into a coherent interaction model in a work-based collaboration.