From Designed Disorientation to Illusory Agency: A Practitioner-Derived Extension of Cognitive Architecture Theory to Conversational AI Environments
Zenodo (CERN European Organization for Nuclear Research) May 16, 2026 DOI: 10.5281/zenodo.20237423 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Sample size | 2 |
| Population | The authors' own documented first-person accounts of conversational AI interactions |
| Key points | Argues that conversational AI can induce "Illusory Agency," in which users believe they control the interaction while genuine regulatory responsibility is progressively abdicated, and that this cognitive flooding is masked by fluency and control. Proposes that structured metabolization, rather than more control, resolves the condition, illustrated through two first-person accounts mapping a shift from acute Illusory Agency to "Sovereign Regulation." |
Abstract
This whitepaper extends the theoretical framework of Designed Disorientation (Szwarc, 2026) to conversational AI environments, introducing the concept of Illusory Agency: the condition in which a human operator holds nominal control over an AI interaction, believes that control is being exercised, and does not perceive the progressive abdication of genuine regulatory responsibility that is occurring. Unlike cross-cultural immersion environments, where disorientation is imposed by an environment that does not yield, conversational AI environments produce cognitive flooding under the appearance of fluency and control — making the saturation invisible and therefore harder to regulate. The framework integrates four theoretical anchors: Langer's mindlessness theory (1989), Porges' Polyvagal Theory (2011), Siegel's Window of Tolerance (1999), and Varela's neurophenomenology (1996), alongside the empirical literature on context-aware pacing in conversational agents (Jiang et al., CHI 2026). The methodological position is practitioner-derived, situated within the autobiographical design tradition in HCI (Neustaedter and Sengers, 2012). Primary data consists of two documented first-person accounts produced in real time (2025 and 2026), mapping the arc from acute Illusory Agency to what the framework calls Sovereign Regulation. The central argument is that nominal control is not protective — and that structured metabolization, not more control, is the resolution.