An exploratory semantic clustering study of how people describe AI art: Implications for psychotherapy
Year: 2026
Type: article
Study design: Observational Study
Abstract: The emergence of Artificial Intelligence (AI) art challenges established aesthetic frameworks, frequently resulting in a documented negative bias when non-human authorship is known. This exploratory study investigated the cognitive schemas activated by traditional visual art and AI art, and how these mental representations are shaped by individual attitudes towards AI. Using the free association technique, 560 participants provided three descriptive words for both art forms, which were subsequently quantified using word embeddings. Principal Component Analysis (PCA) revealed a significant semantic distinction between the vocabularies used for the two art types, with specific clusters emerging for both art forms. These clusters emphasize that traditional visual art is thought of in terms of emotionality and expressiveness, the creative process, aesthetic beauty, color, and art forms and styles which can express the aforementioned dimensions, while AI art is centered on technological innovation, novelty, artificiality, lack of emotion and authenticity, and themes related to fake and stolen content. Additionally, positive attitudes toward AI were associated with greater use of positive AI-related words and reduced use of negative terms. Alternatively, positive AI attitudes negatively predicted the use of traditional art words related to emotion and expressiveness and the creative process. These findings show that AI art activates a more ambivalent and technologically oriented schema compared to the predominantly positive, emotionally grounded schema of traditional art. We discuss these findings in terms of their implications for psychotherapy and how AI-generated imagery may differ from traditional creative expression in therapeutic contexts.
Source: New Ideas in Psychology
Institutions: Alexandru Ioan Cuza University, Sapienza University of Rome
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: hybrid
Estimated APC paid: $3,600
OpenAlex ID: W7162565950