A recent study conducted by Bowling Green State University unveils that generative artificial intelligence (AI) can create images that make it difficult for individuals to identify whether they were produced by an AI or a human artist. However, the study also found that people generally have a subconscious preference for human-made art, even if they cannot provide a clear explanation as to why.
Andrew Samo, a doctoral candidate studying industrial and organizational psychology at BGSU, collaborated with Distinguished Research Professor Dr. Scott Highhouse on this research. Their findings, published in the journal Psychology of Aesthetics, Creativity, and the Arts, challenge the notion that art is uniquely human, as AI-generated art can elicit similar emotional responses. However, individuals still exhibited a stronger affinity for human art.
The research aimed to eliminate bias by not informing participants that they would be viewing both AI and human-generated art. Participants were unaware that some of the artwork was created by AI and were asked to rate a series of pictures based on aesthetic judgment factors. This approach allowed the researchers to determine whether individuals could distinguish between AI art and human art without any deceptive influence.
The findings revealed that participants were able to correctly identify the source of the artwork only slightly more than half the time, and they were not confident in their guesses. There was a 50% to 60% chance of participants correctly identifying the origin of the art, and even when they were able to do so, their confidence levels were low. This difficulty in differentiation was accompanied by the discovery that people had a consistent preference for human artwork, despite not being able to explain the reasons behind this preference.
Upon analyzing the data, Samo and Highhouse identified clear differences in individuals’ emotional responses towards AI art compared to human art. Although participants could not confidently differentiate between the two, they consistently expressed more positive emotions when viewing human-generated art. Factors that contributed significantly to this preference included self-reflection, attraction, nostalgia, and amusement. These results imply that people may feel a stronger connection to human art, even if they are unable to articulate why.
Interestingly, when asked to explain their preference, participants could not provide a clear explanation. One interpretation put forth by the researchers is that snap judgments made a subconscious connection with human art, while analytical processing failed to articulate the underlying reasons for this preference.
The researchers also considered the possibility that the human brain may detect subtle differences in art produced by AI. They suggested that the uncanny valley effect, in which something looks human but has slight perceptual discrepancies that the subconscious can detect, could be responsible for these preferences.
Generative AI has come a long way, surpassing initial expectations that it could only replicate repetitive and routine tasks. With the development of generative AI models, the technology can now create art, music, poetry, prose, and text that is almost indistinguishable from human creations. This advancement presents exciting possibilities for the future applications of generative AI.
As generative AI models continue to improve and become more widely available, it is crucial to understand the psychological effects and human impacts of AI. The study conducted by Samo and Highhouse provides valuable insights into the evolving capabilities of AI and its potential influence on the art world. It also calls for continued research in this field as AI models become increasingly powerful and integrated into everyday life.
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it