Beyond One-to-One Targeted Messaging: A Unified Recipient-Scoped Visibility Architecture for Multi-Participant Generative AI Conversations

The emergence of multi-participant generative artificial intelligence environments is transforming conversational systems from conventional human–AI interactions into collaborative communication spaces involving multiple human participants, large language models (LLMs), autonomous AI agents, and heterogeneous computational services. Although conventional messaging platforms provide private conversations, group conversations, and, increasingly, targeted messages, existing communication paradigms do not fully address the need for a unified, general-purpose mechanism through which the visibility of an individual message can be independently determined by its sender at the time of transmission. In particular, a message intended for one participant should not automatically become accessible to every other participant merely because they share the same conversational environment. This paper proposes Universal Recipient-Scoped Visibility Control (URVC), a conceptual architecture for controlling message-level visibility in multi-participant generative AI conversations. The central principle is that message membership in a shared conversation and message visibility are independent properties. Each message is associated with an explicitly defined recipient set, and access is granted exclusively to authorized recipients rather than automatically inherited from the membership of the surrounding conversation. The proposed framework generalizes recipient-specific communication beyond conventional one-to-one human–agent interactions to support multiple designated human recipients, multiple AI models, autonomous agents, and mixed human–AI recipient groups. The architecture introduces a recipient-scoped message representation, an independent visibility policy, an authorization enforcement layer, recipient-aware context construction, and controlled information propagation. It further establishes formal confidentiality objectives, including recipient isolation, unauthorized-context exclusion, attachment-level protection, and resistance to indirect information leakage through AI-generated summaries and cross-model communication. Unlike an interface-only approach that merely hides messages from unauthorized participants, URVC treats visibility as an enforceable property of the underlying information-processing architecture. To the best of our knowledge, this paper proposes one of the first unified conceptual architectures specifically directed at generalizing recipient-scoped visibility from targeted human–agent messaging into a multi-recipient, multi-model, human–AI conversational environment. The work does not claim to invent targeted messaging itself. Instead, it identifies and formalizes a broader architectural direction for making message-level recipient authorization a fundamental capability of generative AI communication systems.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-16
DOI
https://doi.org/10.5281/zenodo.24269343
Primary Topic
Access Control and Trust
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Beyond One-to-One Targeted Messaging: A Unified Recipient-Scoped Visibility Architecture for Multi-Participant Generative AI Conversations

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

Beyond One-to-One Targeted Messaging: A Unified Recipient-Scoped Visibility Architecture for Multi-Participant Generative AI Conversations

Jincheng Zhang
preprint en

Abstract

The emergence of multi-participant generative artificial intelligence environments is transforming conversational systems from conventional human–AI interactions into collaborative communication spaces involving multiple human participants, large language models (LLMs), autonomous AI agents, and heterogeneous computational services. Although conventional messaging platforms provide private conversations, group conversations, and, increasingly, targeted messages, existing communication paradigms do not fully address the need for a unified, general-purpose mechanism through which the visibility of an individual message can be independently determined by its sender at the time of transmission. In particular, a message intended for one participant should not automatically become accessible to every other participant merely because they share the same conversational environment. This paper proposes Universal Recipient-Scoped Visibility Control (URVC), a conceptual architecture for controlling message-level visibility in multi-participant generative AI conversations. The central principle is that message membership in a shared conversation and message visibility are independent properties. Each message is associated with an explicitly defined recipient set, and access is granted exclusively to authorized recipients rather than automatically inherited from the membership of the surrounding conversation. The proposed framework generalizes recipient-specific communication beyond conventional one-to-one human–agent interactions to support multiple designated human recipients, multiple AI models, autonomous agents, and mixed human–AI recipient groups. The architecture introduces a recipient-scoped message representation, an independent visibility policy, an authorization enforcement layer, recipient-aware context construction, and controlled information propagation. It further establishes formal confidentiality objectives, including recipient isolation, unauthorized-context exclusion, attachment-level protection, and resistance to indirect information leakage through AI-generated summaries and cross-model communication. Unlike an interface-only approach that merely hides messages from unauthorized participants, URVC treats visibility as an enforceable property of the underlying information-processing architecture. To the best of our knowledge, this paper proposes one of the first unified conceptual architectures specifically directed at generalizing recipient-scoped visibility from targeted human–agent messaging into a multi-recipient, multi-model, human–AI conversational environment. The work does not claim to invent targeted messaging itself. Instead, it identifies and formalizes a broader architectural direction for making message-level recipient authorization a fundamental capability of generative AI communication systems.

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