AI Image Generators and Art: What Makes Human Art Valuable?
What makes art valuable? Creatives, museum-goers, and web-surfers have asked this question for ages; whenever a new art form emerges, people reevaluate the answer. Today, generative AI poses unique questions about artistic legitimacy. Some argue that people should accept AI-created art as the inevitable future of art. However, the pushback against AI art differs from skepticism about other mediums. This is because the algorithms that generate images train on people’s photos and artwork without obtaining their consent or crediting them (Kircher and Holtermann). The generation process results in plagiarized images that AI prompters can claim to own and sell. As a result, unethically-sourced AI generators harm creatives by denying them rightful recognition and compensation for their work.
According to NBC News, “Image synthesis models like Google Images, DALL-E and Stable Diffusion are trained using datasets of millions of images. The models learn associations between the arrangement of pixels in an image and the image’s metadata, which typically includes text descriptions of the image subject and artistic style” (Sung). Datasets can include photos of copyrighted work, as well as images people have posted for their followers on social media. Some argue that the AI training process is equivalent to a human artist drawing inspiration from a variety of influences to make original work. However, this is not an accurate comparison, as computer algorithms cannot creatively innovate. As explained above, they recognize the relationships between pixel arrangements and text. An algorithm cannot understand the subjective ideas behind an artwork; it can merely piece together data from different sources to fit a text description. Consequently, AI-generated products may look incredibly similar to existing works.
Human artists spend their lives developing distinct styles and thoughtful perspectives to create meaningful pieces. Lifting work from artists without compensating them devalues their time, effort, and skill. Moreover, when a human artist is heavily inspired by or directly reproduces parts of others’ artworks, they are expected to cite their influences; why would dataset collectors be exempt from this responsibility? Recently, a coder filed a class-action lawsuit against an AI code generator called GitHub Copilot for “infringing copyright because it does not provide attribution when Copilot reproduces open-source code covered by a license requiring it” (Knight). The outcome of this lawsuit could impact the direction of all sorts of generative AI tools. Art generators might become more ethical if developers limit the contents of their datasets to willingly submitted files.
The issue with treating broadly sourced generative AI like any other artistic medium becomes more apparent when you consider instances like this one: “[A Colorado resident] got the idea to submit one of his Midjourney creations to the Colorado State Fair, which had a division for ‘digital art/digitally manipulated photography.’ . . . He had won the division, along with a $300 prize” (Roose). Like other apps mentioned, Midjourney relies on input images from the open web. This contestant entered a text prompt, selected an image from a few results, and printed it out to submit to an art fair. In doing so, he claimed ownership of the image; this allowed him to profit from a machine’s seconds of work to the direct detriment of passionate, practiced artists. Affording such a high level of legitimacy to AI-generated art means disregarding humanity’s role in artistic processes. Unlike artists, prompt inputters leave creative thought, intention, and execution up to an uninventive algorithm. Semi-stolen images that people can mass-produce with machines should not qualify as art because such a framework diminishes the emotional and thematic aspects of art. Disparaging artists’ work causes them to lose necessary financial support.
Fortunately, creatives can still use the internet to thrive. Social media sites are great places for artists to build audiences for their work. It is easy and fun to scroll through Instagram and like, share, and follow creative projects. Online inspiration and interactions can help young artists build their own skills and style. In the end, people stay creative by expressing human thoughts; everyone has artistic potential, and that is truly valuable.
Works Cited
Kircher, Madison Malone, and Callie Holtermann. “How Is Everyone Making Those A.I. Selfies?” The New York Times, 7 Dec. 2022, www.nytimes.com/2022/12/07/style/lensa-ai-selfies.html.
Knight, Will. “This Copyright Lawsuit Could Shape the Future of Generative AI.” WIRED, 21 Nov. 2022, www.wired.com/story/this-copyright-lawsuit-could-shape-the-future-of-generative-ai/.
NBC News. 6 Dec. 2022, www.nbcnews.com/tech/internet/lensa-ai-artist-controversy-ethics-privacy-rcna60242.
Roose, Kevin. “An A.I.-Generated Picture Won an Art Prize. Artists Aren’t Happy.” The New York Times, 2 Sept. 2022, www.nytimes.com/2022/09/02/technology/ai-artificial-intelligence-artists.html.
