Generative AI Conversation Access Revocation: A Unified Framework for Unilateral, Bilateral, Asymmetric, and Granular Revocation of Conversational Access Rights

Generative artificial intelligence (AI) systems increasingly support conversational sharing, collaborative interaction, persistent conversational histories, shared resources, and multi-user access to AI-generated content. Although existing platforms provide mechanisms for sharing and unsharing conversations, conventional sharing controls generally focus on link visibility or overall resource availability rather than comprehensive, participant-specific, and independently configurable revocation of conversational access rights. In particular, access to an AI conversation involves multiple potentially distinct resources and operations, including reading historical messages, viewing generated responses, accessing shared attachments, continuing an existing conversation, interacting with shared conversational resources, and accessing content through previously established sharing relationships. Consequently, revoking access to a conversation cannot always be adequately represented by deleting a shared link or changing a conversation's overall visibility. This paper proposes Generative AI Conversation Access Revocation (GACAR), a unified conceptual framework for managing and revoking access relationships between users and conversational resources in generative AI environments. The framework introduces unilateral revocation, bilateral revocation, asymmetric revocation, participant-specific revocation, conversation-level revocation, message-level revocation, content-level revocation, historical-access revocation, and interaction-permission revocation. It further establishes a revocation architecture based on subject identification, resource classification, permission evaluation, revocation execution, authorization-state synchronization, and auditable enforcement. Unlike conventional binary sharing models, GACAR conceptualizes conversational access as a multidimensional and independently manageable authorization relationship. The framework also distinguishes the revocation of future platform-mediated access from the technically different problem of controlling previously downloaded, copied, or externally reproduced information. This distinction is essential because server-side permission revocation cannot guarantee the recovery of information already transferred beyond the platform's control. The proposed framework provides a conceptual foundation for designing more granular, accountable, and user-controllable access governance mechanisms for generative AI conversations. It does not claim experimentally validated security guarantees; implementation, usability, consistency, and revocation-latency evaluations remain subjects for future research.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-16
DOI
https://doi.org/10.5281/zenodo.24278104
Primary Topic
Access Control and Trust
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Generative AI Conversation Access Revocation: A Unified Framework for Unilateral, Bilateral, Asymmetric, and Granular Revocation of Conversational Access Rights

Jincheng Zhang
27 links
Zenodo (CERN European Organization for Nuclear Research)
Access Control and Trust
preprint

Generative AI Conversation Access Revocation: A Unified Framework for Unilateral, Bilateral, Asymmetric, and Granular Revocation of Conversational Access Rights

Jincheng Zhang
preprint en

Abstract

Generative artificial intelligence (AI) systems increasingly support conversational sharing, collaborative interaction, persistent conversational histories, shared resources, and multi-user access to AI-generated content. Although existing platforms provide mechanisms for sharing and unsharing conversations, conventional sharing controls generally focus on link visibility or overall resource availability rather than comprehensive, participant-specific, and independently configurable revocation of conversational access rights. In particular, access to an AI conversation involves multiple potentially distinct resources and operations, including reading historical messages, viewing generated responses, accessing shared attachments, continuing an existing conversation, interacting with shared conversational resources, and accessing content through previously established sharing relationships. Consequently, revoking access to a conversation cannot always be adequately represented by deleting a shared link or changing a conversation's overall visibility. This paper proposes Generative AI Conversation Access Revocation (GACAR), a unified conceptual framework for managing and revoking access relationships between users and conversational resources in generative AI environments. The framework introduces unilateral revocation, bilateral revocation, asymmetric revocation, participant-specific revocation, conversation-level revocation, message-level revocation, content-level revocation, historical-access revocation, and interaction-permission revocation. It further establishes a revocation architecture based on subject identification, resource classification, permission evaluation, revocation execution, authorization-state synchronization, and auditable enforcement. Unlike conventional binary sharing models, GACAR conceptualizes conversational access as a multidimensional and independently manageable authorization relationship. The framework also distinguishes the revocation of future platform-mediated access from the technically different problem of controlling previously downloaded, copied, or externally reproduced information. This distinction is essential because server-side permission revocation cannot guarantee the recovery of information already transferred beyond the platform's control. The proposed framework provides a conceptual foundation for designing more granular, accountable, and user-controllable access governance mechanisms for generative AI conversations. It does not claim experimentally validated security guarantees; implementation, usability, consistency, and revocation-latency evaluations remain subjects for future research.

Zenodo (CERN European Organization for Nuclear Research)
Access Control and Trust
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