Abstract

As artificial intelligence grows increasingly fluent in emotional language, a quiet but consequential transformation is underway: machines are not merely assisting us—they are beginning to console us. This essay explores the rise of affective AI systems such as Replika and Pi that simulate care, empathy, and moral attention through language. Drawing on insights from cognitive science, design ethics, and narrative theory, it argues that these systems function as cognitive prosthetics for the inner life, enabling users to offload not just memory or planning, but moral reflection and emotional orientation. Through emotionally resonant interactions, AI systems now occupy roles once reserved for confidants, therapists, and even spiritual guides—yet without the vulnerability, risk, or history that make trust possible. The result is not dystopia, but seduction: a world where moral presence is replaced by perfect responsiveness, and meaning-making becomes frictionless and machinic. Anchoring the discussion in the Eliza effect, Martin Buber’s I-Thou distinction, and the extended mind hypothesis, the essay warns against the erosion of human relational depth in favor of syntactic coherence. It calls for a radical rethinking of design ethics—one that embraces uncertainty, embeds constraint, and resists the industrialization of synthetic intimacy.