Answers are grounded in this paper and cite their sources.
An AI response can be accurate, useful, even extraordinary. Its quality alone cannot establish what draws a person back or what repeated interaction sustains, changes, or makes possible. MWe propose Human–AI Coupling as a developmental framework for examining how recurrent, contingent interaction with a deployed AI service enters a person's judgment, regulation, work, and participation across time. The focal unit is a person–service–function–time relationship. Seven observational dimensions—cognitive, creative, emotional, embodied, relational, behavioral, and ecological—organize observation; nine research propositions examine possible pathways to later change. Four outcome distinctions separate immediate experience, assisted performance, sustained supported functioning, and retained or transferable capability. The program investigates newly feasible projects, reliable access, anticipation, attachment, repair work, and restricted choices, allowing beneficial, harmful, neutral, or mixed outcomes. Within Interrasyntherics, it connects developmental, cultural, and ecological analysis. MWe further propose Human–AI Dyadic Idiolect to study linguistic conventions whose meanings and uses develop through recurring exchange. Proposed tests compare explanations and decisions with equally informed existing approaches, accounting for the work required. This is a conceptual synthesis and prospective research program; no new participant study is reported. This work was completed independently and received no funding. The author declares no competing interests.
No takes yet. Share an insight, caveat, or question.
Bia (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: