September 2015
Volume 15, Issue 12
Free
Vision Sciences Society Annual Meeting Abstract  |   September 2015
The artistic Turing test: An exploration of perceptions of computer-generated and man-made art
Author Affiliations
  • Rebecca Chamberlain
    KU Leuven
  • Caitlin Mullin
    KU Leuven
  • Johan Wagemans
    KU Leuven
Journal of Vision September 2015, Vol.15, 112. doi:https://doi.org/10.1167/15.12.112
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Rebecca Chamberlain, Caitlin Mullin, Johan Wagemans; The artistic Turing test: An exploration of perceptions of computer-generated and man-made art. Journal of Vision 2015;15(12):112. https://doi.org/10.1167/15.12.112.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

The generation of genuinely creative works of art could be considered as the final frontier in artificial intelligence (AI). Several AI research groups are pursuing this by programming algorithms which generate works in various media and styles. The ultimate test of success for such a machine artist would be to convince the onlooker that it was generated by a human being; an artistic Turing test. Previous research has shown that observers can distinguish artworks generated by a skilled human artist over those of a child or an animal using the perception of intentionality- the appearance of a planned final product (Hawley-Dolan & Winner, 2011). However, there is a little research on whether observers can differentiate between computer-generated art and man-made art, and if so, whether these judgments are driven by impressions of intentionality or surface characteristics that identify the mode of production. Furthermore, the decision that an artwork is computer-generated may reflect a negative aesthetic preference, as research has suggested that believing an artwork or musical composition to be computer-generated negatively affects its aesthetic appraisal (Kirk et al, 2009; Moffat & Kelly, 2006). The current study examined whether individuals were able to differentiate between works of art whose creative or representational abilities are defined by computer algorithms, from matched artworks created by human artists. Participants sorted artworks into computer- or human-generated and indicated their aesthetic preference on a 7 point Likert-scale. Results show that participants were able to successfully determine the provenance of the artworks. Perception of intentionality and surface characteristics as well as a subset of image statistics (Pyramid of Histograms of Orientation Gradients (PHOG), luminance spectra) were also investigated in relation to source decision criteria. The results have implications for the way in which AI algorithmic art is created, as well as providing insights into their aesthetic perception.

Meeting abstract presented at VSS 2015

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×