Generative AI Interruption Permission: The First Unified Directional and Configurable Permission Framework for Interruption and Supplementation in Human–AI Conversations

Human–AI conversations have traditionally been organized around turn-taking mechanisms in which participants exchange speaking opportunities according to predefined interaction patterns. Although previous research has investigated user-initiated interruption, system-initiated interruption, speech overlap, and conversational turn management, interruption is predominantly treated as a technical event or a dialogue-management problem rather than as an independently configurable interaction permission. This paper introduces Generative AI Interruption Permission (GAIP), a unified conceptual and architectural framework that models interruption and conversational supplementation as explicitly configurable permissions between generative AI systems and human participants. To the best of the author's knowledge, this is the first unified framework specifically proposed to formalize directional, configurable, and revocable interruption and supplementation permissions as an independent permission-management layer for generative AI conversations. GAIP distinguishes between user-to-AI interruption, AI-to-user interruption, bidirectional interruption, and unidirectional interruption, while extending the model to multi-participant environments in which permissions can differ according to the identity of the interrupter and the recipient. The framework further introduces permission scope, authorization status, conditional activation, temporary authorization, revocation, interruption priority, supplementation rights, conversational recovery, and permission inheritance as core design dimensions. Rather than assuming that every detected interruption should automatically result in a change of speaking control, GAIP separates the detection of an attempted interruption from its authorization and execution. This separation provides a foundation for more predictable, controllable, and transparent conversational interaction. The paper develops the conceptual foundations, formal permission model, architectural components, interaction lifecycle, representative configurations, governance requirements, and an empirical evaluation agenda. GAIP contributes a new perspective in which conversational interruption is understood not merely as an acoustic or linguistic phenomenon but as a permission-governed interaction capability.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-17
DOI
https://doi.org/10.5281/zenodo.24297673
Primary Topic
Speech and dialogue systems
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Generative AI Interruption Permission: The First Unified Directional and Configurable Permission Framework for Interruption and Supplementation in Human–AI Conversations

Jincheng Zhang
26 links
Zenodo (CERN European Organization for Nuclear Research)
Speech and dialogue systems
preprint

Generative AI Interruption Permission: The First Unified Directional and Configurable Permission Framework for Interruption and Supplementation in Human–AI Conversations

Jincheng Zhang
preprint en

Abstract

Human–AI conversations have traditionally been organized around turn-taking mechanisms in which participants exchange speaking opportunities according to predefined interaction patterns. Although previous research has investigated user-initiated interruption, system-initiated interruption, speech overlap, and conversational turn management, interruption is predominantly treated as a technical event or a dialogue-management problem rather than as an independently configurable interaction permission. This paper introduces Generative AI Interruption Permission (GAIP), a unified conceptual and architectural framework that models interruption and conversational supplementation as explicitly configurable permissions between generative AI systems and human participants. To the best of the author's knowledge, this is the first unified framework specifically proposed to formalize directional, configurable, and revocable interruption and supplementation permissions as an independent permission-management layer for generative AI conversations. GAIP distinguishes between user-to-AI interruption, AI-to-user interruption, bidirectional interruption, and unidirectional interruption, while extending the model to multi-participant environments in which permissions can differ according to the identity of the interrupter and the recipient. The framework further introduces permission scope, authorization status, conditional activation, temporary authorization, revocation, interruption priority, supplementation rights, conversational recovery, and permission inheritance as core design dimensions. Rather than assuming that every detected interruption should automatically result in a change of speaking control, GAIP separates the detection of an attempted interruption from its authorization and execution. This separation provides a foundation for more predictable, controllable, and transparent conversational interaction. The paper develops the conceptual foundations, formal permission model, architectural components, interaction lifecycle, representative configurations, governance requirements, and an empirical evaluation agenda. GAIP contributes a new perspective in which conversational interruption is understood not merely as an acoustic or linguistic phenomenon but as a permission-governed interaction capability.

Zenodo (CERN European Organization for Nuclear Research)
Speech and dialogue systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon