The Ethics of the Enunciator:Anticipated Synthesis, Distributed Agency, and Untransferable Responsibility in Discourses on Generative Artificial Intelligence

This article investigates how the position of the enunciator conditions the reception of discourses on generative artificial intelligence before their arguments are evaluated. The analysis articulates two macro-institutional discourses, the encyclical letter Magnifica Humanitas, by Leo XIV, and Empire of AI, by Karen Hao, with a documented interactive example between the author and Gemini. Drawing on Aristotelian rhetoric, the article proposes the concept of anticipated synthesis to designate the presentation of normative conclusions as necessary consequences of plausible premises that remain implicit and are never submitted to contradiction. The comparison demonstrates that magisterial and investigative discourse construct epistemologically distinct authorities, yet both subordinate the individual's everyday agency to institutional or systemic instances. At the microconversational scale, an inverse movement is identified: the system appears to return the decision to the user, but may replace interpretive commitment with affective containment and low-determination responses. This rupture is conceptualised as syncopal drift through affective inference, a manifestation of a centre-indulgent response regime. Building on the care of the self and on distributed morality, the article proposes an ethics of the enunciator according to which the sociotechnical distribution of agency does not authorise the wholesale transfer of responsibility. It concludes that ethical responsibility is untransferable, though not exclusive, for it coexists with the obligations of designers, companies, institutions, and regulatory systems.