This conceptual paper explores whether artificial systems could emulate emotion-like processes by integrating affective cues with memory and action selection. It argues that emotional expression and consciousness are conceptually separable and examines the ethical implications of affective but non-conscious AI systems. Through a speculative architectural account, the paper discusses how affect could support adaptive decision-making in complex environments while remaining distinct from sentience. The work is relevant to emotional AI, ethical evaluation, and the limits of anthropomorphic interpretations of AI behavior.

Author of the paper: Hermann Borotschnig

Publisher or journal of publication: AI & Society

The paper is available at the following link.