How do humans perceive art created by Artificial Intelligence?

“Artificial intelligence” (AI) is ubiquitous in our everyday lives these days. While the technology is incorporated not only into smartphones, translators, voice assistants and self-driving cars, it has now also pathed its way into the art world. For instance, AI is able to recreate paintings of well-established artists (Iansiti & Lakhani, 2020), but can generate original art styles (Schwab, 2017), songs (Vincent, 2016), or poems as well (Gibbs, 2016). It is usually impossible for people to distinguish between human-made and AI-created art, hence, they often place high artistic (Elgammal et al., 2017) as well as monetary value on AI artwork (BBC, 2018). A recent study titled “Defending humankind: Anthropocentric bias in the appreciation of AI art” published in Computers in Human Behavior investigated how people react to art created by AI systems and labeled as such, compared to artwork labeled as human-made.

Click here to access the original publication.

Research overview

The study was conducted by Dr. Florian Buehler, who is senior lecturer at Vorarlberg University of Applied Science and a guest lecturer at HfT Stuttgart, and his colleagues from the Vrije Universiteit Amsterdam and University of British Columbia. It comprised of four experiments, totaling 1,708 participants. While the first experiment surveyed university students in the Netherlands, participants in the following experiments were UK residents.

Study participants rated two artworks: one “AI-made” and one “human-made”. Labels were randomized to assess whether there was a bias against AI-made art regardless of the presented content. One experiment utilized two short music clips created by AIVA. Another experiment used paintings by famous artists (Paul Klee and Kandinsky), while in yet another experiment participants were shown images created by the autonomous AI artist AICAN. Posters of artworks served as stimuli in the last experiment. Following exposure to the music (paintings, or posters), participants were then asked which of the two artworks they admired more in respect to the emotional experience of awe and which they thought was more creative.

Main findings

  • Participants experienced less awe for art labeled as AI-made (vs. human-made).
  • Participants perceived art labeled as AI-made (vs. human-made) as less creative.
  • Anthropocentric, meaning the perceived precedence of humans over other species, creativity beliefs moderate the effect on awe. More specifically, the negative perception towards AI-generated art was found only for participants who scored high on anthropocentric creativity beliefs.
  • Participants would be less likely to buy AI-generated art than human-made art, which was also more pronounced among people who score high on anthropocentric creativity beliefs.
  • The following serial mediation effect was found: Exposure to AI-made (vs. human-made) art -> perceived as less creative -> less awe -> lower preference to buy. This effect again held more for people with stronger anthropocentric creativity beliefs.


This research concluded that art generated by AI is perceived as less creative and inspiring compared to human-created artwork. This perception is especially true for people who believe that creativity is a uniquely human characteristic. The results therefore suggest that there is a distortion of perception when artwork is created by AI, possibly because the participants experienced an ontological threat to the human view on creativity. In other words, our brains typically associate artistic creativity with human intelligence. However, when we realize that AI can also create art, our ideas about creativity and intelligence are shaken. Overall, this study illustrates the need to rethink our ideas about the definition of creativity and who can be creative. AI presents opportunities to enhance our human capabilities rather than threaten them.


BBC News. (2018). Portrait by AI program sells for $432,000. BBC News. October 25

Elgammal, A., Liu, B., Elhoseiny, M., & Mazzone, M. (2017). Can: Creative adversarial networks, generating “art” by learning about styles and deviating from style norms. arXiv preprint arXiv:1706.07068.

Gibbs, S. (2016). Google AI project writes poetry which could make a vogon proud. The Guardian, Guardian News and Media, 17 May 2016

Iansiti, M., & Lakhani, K. R. (2020). Competing in the age of AI: Strategy and leadership when algorithms and networks run the world. Harvard Business Press.

Millet, K., Buehler, F., Du, G., & Kokkoris, M. D. (2023). Defending humankind: Anthropocentric bias in the appreciation of AI art. Computers in Human Behavior, 143, 107707.b

Schwab, K. (2017). The fourth industrial revolution. Currency.

Vincent, J. (2016). This AI-written pop song is almost certainly a dire warning for humanity. Verge.