A recent arXiv paper (2610.11410) examines how affective computing is evolving. According to the abstract, the field has progressed from recognizing discrete emotion categories toward more open-ended affective analysis powered by large multimodal models. This marks a technical shift in how AI systems attempt to interpret human emotion.
The paper then contrasts this trajectory with affective science, which frames emotion not as a static label but as an unfolding process influenced by appraisal, regulation, and other factors. The abstract suggests that current AI approaches may not fully capture this dynamic nature, though the full details of the proposed framework are not available from the abstract alone.
Because the source is only an abstract, the article cannot describe specific methods or results. What is clear is that the authors are calling attention to a conceptual gap between engineering practice and psychological theory in emotion AI.