Liking Without Endorsing: Consumer Dilemmas in Responses to AI‐Generated Music

ABSTRACT Despite a growing literature documenting consumer aversion to AI‐generated creative output, AI‐generated music has achieved notable success in the marketplace, with some songs achieving chart placements and attracting millions of streams. We reconcile this discrepancy by identifying music as a uniquely affect‐driven art form in which authorship cues carry less weight in quality judgments following direct content exposure. Across three experiments, a pilot study, and multiple supplemental tests, we show that although consumers initially report lower expected enjoyment for AI‐generated (vs. human‐generated) music, these differences attenuate following direct exposure. We further demonstrate that this pattern is domain‐specific: Whereas evaluations of AI‐generated music converge with those of human‐generated music after listening, penalties for AI authorship persist in more cognitively mediated domains (e.g., writing). Importantly, such a convergence in aesthetic evaluation does not reflect unqualified acceptance. Instead, consumers experience evaluative conflict as they reconcile favorable listening experiences with prior aversive beliefs about AI creativity. Further, despite increased enjoyment, consumers remain less willing to engage deeply with AI‐generated music through behaviors like relistening and sharing. Together, our findings provide a novel account of how consumers evaluate AI‐generated creativity, highlighting the role of affective processing in attenuating—but not fully resolving—AI authorship concerns in music.