Recent AI-centered scientific news reports often undergo rapid narrative transformation during downstream dissemination. This study proposes a preliminary structural model—the Three-Speed Translation Model—to analyze such translation dynamics. Narrative transformation is operationalized using three independent indices: Lexical Heat Index (LHI), Subject Substitution Ratio (SSR), and Primary Link Salience (PLS). Using ordered publication layers (Primary Scientific Source → Corporate Communication → Media Coverage → Commentary), we apply Spearman rank analysis to examine directional translation patterns. Preliminary results suggest asymmetric dynamics: LHI increases across downstream layers, SSR peaks at the corporate layer before stabilizing, and PLS decreases progressively. Mythologization is operationally defined as a measurable increase in lexical thermalization combined with decreased primary-source salience across ordered publication layers. This pilot study establishes a structural and measurable framework for analyzing technological myth formation and invites further cross-domain validation.
Hinano Kimura (Sun,) studied this question.