Get started
Work

Generated by Creative AI: Exploring Human Responses to Information About AI as the Source of Creative Content

Year: 2025
Type: dissertation
Abstract: Artificial intelligence (AI) can now be used to produce a wide range of creative content, such as images, stories, music or poems. Reactions to this technological advancement could have an impact on the future development and use of this technology. This thesis examines how the information that creative content has been generated by an AI influences the experience and evaluation of creative works. The focus here is on the reception of stories and images. Possible effects on different forms of experience and judgment are examined using different stimuli materials. In addition, the mechanisms underlying these effects are tested using empirical methods. Furthermore, the potential influence of various individual differences on reactions to AI artist or author information is examined. Manuscript 1 presents the results of two experiments (N1 = 325, N2 = 489) that indicate a negative influence of AI author information on transportation into a story. Furthermore, it was shown that genre of the story can influence this effect. The negative effects of AI author information are only evident for contemporary fiction stories, but not for the science fiction genre. The eeriness experienced while reading and the willingness to share the story are not influenced by AI artist information as originally hypothesized. Openness, affinity for technology and attitude towards AI do not influence the effect of author information on transportation. Manuscript 2 examines in two experiments (N1 = 176, N2 = 381) whether a lower appreciation of supposedly AI-generated images can be attributed to the reduced agency and experience ascribed to AI compared to humans. Both experiments show a significant indirect effect of AI artist information on the appreciation of AI-generated images, although no total effect can be observed. In addition, a positive deviation from pre-set expectations is considered as a potential mediator of a positive effect of AI artist information on appreciation. However, no mediating effect becomes apparent. Manuscript 3 addresses a possible algorithmic bias in AI-generated images. Results of a content analysis of 324 AI-generated images of people working in STEM professions show that women and POC are strongly underrepresented. Results of an experiment (N = 495) indicate that these biases are not reflected in responses to AI artist information. Rather, the information that images are AI-generated increases the perceived objectivity of some of the images. Furthermore, AI artist information causes moral outrage when the AI-generated images show older people. In sum, the information that creative content has been generated by an AI, and not created by a human, can impact responses to these works. However, these effects can be influenced by the genre or the specific content. The reactions to AI artist or author information do not always have to reflect actual tendencies but depend on perceptions of AI.
Language: English
Open Access status: green
OpenAlex ID: W6964876316
Cites:
Cited by:
Related to: