Interoceptive Divergence in Aesthetic Evaluation and Implications for Human-AI Alignment
The paper discusses the divergence between human and AI aesthetic evaluations. It highlights that while AI can approximate human tendencies in beauty ratings, significant differences remain in emotional responses and bodily sensations. These findings underscore challenges in aligning AI systems with human-like aesthetic processing.
- ▪Artificial intelligence is approaching and sometimes surpassing human performance in cognitive tasks.
- ▪The study compares human and AI responses to aesthetic evaluations using a set of questionnaire items.
- ▪Notable divergences were found in emotional responses and the relationship between beauty ratings and bodily sensations.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Human-Computer Interaction arXiv:2605.18759 (cs) [Submitted on 5 Apr 2026] Title:Interoceptive Divergence in Aesthetic Evaluation and Implications for Human-AI Alignment Authors:Yoshia Abe, Tatsuya Daikoku, Yasuo Kuniyoshi View a PDF of the paper titled Interoceptive Divergence in Aesthetic Evaluation and Implications for Human-AI Alignment, by Yoshia Abe and 2 other authors View PDF HTML (experimental) Abstract:Artificial intelligence (AI), exemplified by large language models (LLMs), is rapidly approaching and in some cases surpassing human performance across a wide range of cognitive tasks. However, human nature is not limited to intelligence alone; it also encompasses sensibility, including the capacity to perceive and experience beauty in visual scenes.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
