This study aims to examine the artistic potential and potential of images created by an artificial intelligence algorithm through 'miscalculation'. GAN is a generative model that is optimized to create fake images similar to the real thing, and StyleGAN is widely used in platforms that provide images of fake people. In this study, images generated due to algorithm errors were collected and then classified into 12 types. The visual effect shown by these images is astonishing and outstanding as it cannot be found in the history of art or photography. In fact, as algorithms intervene in the production of works of art, the issue of creativity in artificial intelligence is a matter of utmost concern. Here, the prevailing view is that algorithms cannot produce creative images due to program’s limitations. Conversely, this study asserts the creative potential of artificial intelligence algorithms based on the innovative results shown by images of miscalculation. At this time, artificial intelligence can expand human vision by providing a new visual experience that has not been experienced in the past.