User-Controlled Read Receipt Disclosure for Generative Artificial Intelligence: A Privacy-Preserving Framework for Selective and Configurable User–AI Read Status Signaling
Year: 2026
Type: preprint
Abstract: Read receipts are widely used in digital communication systems to indicate whether a recipient has viewed a message. Although conventional messaging platforms provide mechanisms for enabling or disabling read receipt visibility, the increasing adoption of generative artificial intelligence (AI) introduces a distinct communication environment in which users interact with AI assistants, conversational agents, and autonomous multi-agent systems. In this environment, the disclosure of user reading activity raises privacy questions that extend beyond conventional interpersonal messaging. Users may wish to access AI-generated messages without informing the AI system, application, or other authorized communication participants that those messages have been read. However, reading activity and permission to disclose reading activity are conceptually different operations and should not automatically be treated as equivalent. This paper proposes User-Controlled Read Disclosure Control for Generative Artificial Intelligence (UCRDC-GAI), a dedicated framework that enables users to independently determine whether, when, and to whom their reading status is disclosed during generative AI interactions. The framework introduces configurable read receipt permissions at system, conversation, AI-agent, and individual-message levels, together with selective, delayed, scheduled, conditional, and temporary disclosure mechanisms. It further establishes a separation between message delivery, presentation, opening, reading-event recording, disclosure authorization, and read receipt transmission. A policy-based architecture is developed to support user authorization, recipient-specific enforcement, privacy transparency, and revocation of future disclosure permissions. The framework also addresses indirect reading-status inference, server-side telemetry, multi-agent information propagation, and the limitations of conventional read receipt controls in AI communication environments. To the best of the author's knowledge, this work presents the first dedicated conceptual framework specifically formulated around independently configurable user-controlled read receipt disclosure for generative AI communication. The contribution is positioned not as the invention of conventional read receipts or their privacy settings, but as the establishment of a distinct AI-oriented disclosure-control model. This research provides a conceptual and technical foundation for implementing privacy-preserving reading-status management in future generative AI communication systems.
Author: Zhang, Jincheng
Language: English
Subfield: Sociology and Political Science
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7234334979
