Answers are grounded in this paper and cite their sources.
Contemporary conversational artificial intelligence systems predominantly organize interaction around user-initiated requests, system-generated responses, turn-taking prediction, interruption management, and automated conversation routing. Although these mechanisms facilitate conversational coordination, they do not adequately conceptualize the proactive transfer of conversational speaking authority as an independent, explicitly controllable interaction operation. This paper introduces Generative AI Speaking-Floor Handoff (GASFH), a unified conceptual framework for enabling a current conversational speaker to proactively transfer speaking authority to another designated participant. The framework encompasses human-to-AI, AI-to-human, human-to-human, and AI-to-AI transfers, supporting both unidirectional and bidirectional configurations. Unlike conventional turn-taking, which primarily concerns the temporal organization of successive contributions, and interruption management, which concerns the disruption or displacement of an ongoing contribution, GASFH conceptualizes the transfer itself as an intentional conversational operation involving an initiating participant, a current speaking-authority holder, a designated recipient, a transfer request or declaration, a transfer condition, and a resulting authority state. The proposed framework further distinguishes transfer initiation from recipient acceptance, actual authority reassignment, and subsequent execution of the recipient's contribution. It introduces a structured lifecycle comprising initiation, validation, notification, acceptance or rejection, commitment, execution, completion, and optional reversal or recovery. Additional capabilities include conditional, scheduled, temporary, persistent, selective, group-based, context-preserving, and priority-aware handoffs. A formal state-transition model, interaction protocol, governance principles, and evaluation framework are developed to support implementation across text-based, spoken, multimodal, multi-agent, and collaborative conversational environments. To the best of our knowledge, this paper presents the first unified conceptual framework specifically dedicated to treating proactive speaking-floor handoff as an independent, directionally configurable, negotiable, and reversible conversational operation across human and generative-AI participants. The contribution is a theoretical and architectural proposal rather than an empirically validated system; experimental validation and comparative evaluation are identified as necessary subsequent research.
No takes yet. Share an insight, caveat, or question.
Jincheng Zhang (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: