Answers are grounded in this paper and cite their sources.
Generative artificial intelligence (GenAI) is increasingly becoming a persistent participant in personal communication, information management, and interpersonal interaction. As AI systems evolve from conventional conversational interfaces into personalized agents, a new requirement emerges: users and AI agents need the ability to maintain contextual information about individual contacts without necessarily disclosing that information to the contacts themselves or to other users and agents. Although conventional contact management applications support manually created contact notes, and contemporary AI systems increasingly incorporate persistent memory, these capabilities do not fully establish a unified model for independently owned, privately visible, and collaboratively managed contact annotations across human and artificial intelligence entities. This paper proposes the Private Contact Notes for Generative Artificial Intelligence (PCN-GAI) framework, a conceptual architecture for managing private annotations associated with individual contacts in multi-user and multi-agent communication environments. The framework distinguishes the identity of a contact from the ownership and visibility of information recorded about that contact. It introduces four primary interaction configurations: user-originated unilateral notes, AI-originated unilateral notes, bidirectional user–AI notes, and jointly maintained notes. In addition, it establishes independent note ownership, contact-specific memory isolation, granular access control, provenance tracking, temporal validity, controlled AI-generated updates, and explicit authorization for information disclosure. Multiple AI systems may maintain separate private notes about the same contact without automatically sharing their contents. The framework further distinguishes the person or entity being described from the owner, creator, editor, and authorized reader of a note. This distinction provides a foundation for preventing unintended information transfer between users, AI agents, and communication participants. The paper formalizes the framework's conceptual entities, operational lifecycle, permission model, security requirements, and evaluation methodology. Rather than claiming experimentally demonstrated performance, it establishes a research and implementation architecture that can be evaluated through controlled experiments, privacy testing, and human–AI interaction studies. PCN-GAI provides a foundation for investigating private, independently controlled, and AI-assisted contact memory as a distinct capability within next-generation generative AI communication systems.
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: