Conceptual framework introduces foundational design principles for psychosocial AI, highlighting distinct engineering approaches for relational human-machine interaction.
Human Psychosocial AI is an emerging field exploring how artificial intelligence can support people through distinct domains of human psychosocial experience. This Working Draft, Part I, establishes the conceptual foundation of the field and positions Human-Centered Relational AI as its first engineering philosophy. It distinguishes Human Psychosocial AI from adjacent approaches including human-centered AI, AI ethics, affective computing, companion AI, mental-health AI, and therapy-oriented AI. The Working Draft proposes four initial principles for Human-Centered Relational AI: continuity, judgment, presence without extraction, and affective intelligence without manufactured intimacy. Each principle is presented with a design implication and an open research question. The paper also proposes an organising logic based on domains of human psychosocial experience rather than industry sectors, illustrated through two current domain specialisations: Appearance Psychosocial AI and Unscripted Life Psychosocial AI. This is an evolving conceptual and research work. It does not claim that the proposed framework or principles have yet been empirically validated. Subsequent parts will extend the framework into further design principles, technical implications, and research questions.
No takes yet. Share an insight, caveat, or question.
Maduewesi et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: