Partners in Conversation

Conversational AI platforms give rise to new modes of human interaction with computers. Whereas command-based interfaces were told what to do by typing lines of texts (in DOS or Linux) or by clicking visual elements (in Macintosh or Windows), Large Language Models (LLMs) rely on intention-based interfaces which respond to users’ prompts about what they want. Far from instantaneous, human and computer often engage in a back-and-forth dialogue over time. They become partners in conversation. In this chapter, Larry Sommer McGrath shows how Chat-GPT, Claude, Duolingo, Gemini, Perplexity, and other conversational AI platforms involve iterative sequences of turn-taking exchanges with users. Each turn has the potential to shift words’ meaning, to follow unforeseen pathways, and to steer the conversation astray in a manner not unlike natural conversation among humans. The chapter suggests that the design of intention-based interfaces can learn from the techniques that people already use in everyday interactions when navigating ambiguity and arriving at mutually desired outcomes. The result is more effective partnerships with conversational AI platforms responsive to diverse styles of social encounters.