AI Art Review: Parallels. AI Art is often questioned regarding…

AI Art is often questioned regarding its training data due to copyright and bias concerns. According to the documentation, Parallels uses “machine learning models that have been trained on the art historical and visual histories of the installation site, as well as other ambient, environmental sensor data.” This potentially means it was trained on artworks by local artists. Maybe the artworks are in the public domain or have licenses like Creative Commons, making the models trained free from copyright questions. The issue of bias is not prominent here, though bias from mainstream society is embedded in mainstream cultural heritage.

Beyond viewing the artwork through the critical lenses of copyright and bias, the screen is part of the environment, and the environment is also reflected in the screen. What makes this piece unique is that, unlike typical AI art, where the focus is on the audience and viewers see themselves transformed through AI, in this case, viewers look at other viewers and the scenery through the artwork. Passersby don’t necessarily engage with the artwork, but it’s interesting to see them through it. Additionally, it’s rare for a screen this large and square-shaped to be placed outdoors. AI art is usually exhibited indoors in an art space. Furthermore, the frame’s minimalism adds a poetic quality.

The documentation of the piece is a well-made 3:50 video, a piece of video art in itself. The narrators articulate philosophical musings about the artwork in contemplative voices. The video encourages us to reflect on how AI enables us to see the world anew through collective cultural heritage.