Example of AI-generated art | photo credit Midjourney
Art and AI are two concepts that should not go hand-in-hand.
Art is defined as humans utilizing natural creativity to create works of varying mediums meant to be viewed for their beauty and emotional power.
Artificial intelligence, or AI, is defined as computers performing tasks normally performed by humans by processing information provided to them.
Yet there’s been a surge of people dubbed “AI artists” appearing across the Internet on sites such as ArtStation and deviantART.
Programs such as DALL·E and Midjourney take user text prompts and use datasets created by millions of images from around the internet to form original art pieces from scratch. While early uses of these AI art generators were primarily for fun and the enjoyment of their users, their rise in popularity has led to individuals selling AI-generated products for profit.
One egregious example of this is the children’s book Alice and Sparkle “written” by Ammaar Reshi. The veracity of Reshi’s authorship is very loose, however, as the book’s written content comes from the AI chatbot ChatGPT and its visual content came from the aforementioned Midjourney. Reshi announced the tools he had used on Twitter and stated that in only a weekend, he went from “ideas, to illustrations, to becoming a published author” through Amazon’s Kindle Direct Publishing.
Since the release of the book on December 4 of last year, Reshi’s book and others produced in a similar manner have faced backlash on Twitter, primarily from actual children’s book illustrators. Adriane Tsai, a children’s book illustrator from Southern California, posted a Tweet stating that even if the AI generation looked like new art, art created from stolen art was still created from stolen art. A later response to Tsai’s Tweet asked why it was okay for artists to look at others’ drawings and take away something from them and why AI couldn’t do the same.
Simple.
Artists take inspiration from others’ works and add their own flair to them to make them their own, whereas AI takes information from others’ works and claims to make them their own.
Professional artist Karla Ortiz, who has over 13 years of experience across dozens of different projects, stated in a post on her blog KOrtiz Blog that AI processes information differently than humans. Her research determined that AI could not interpret inspiration the same way most humans do. They highly depend on the works of others to generate artwork and only what they are told to generate. No unique nuances can be added to this art seeing as a machine can only work with whatever it is given. Ergo, the artwork is but plagiarized pastiches of other artists’ works.
Furthermore, the generation of AI art has dubious ethics and legality behind it.
Since AI datasets pull from across the Internet and even art creation tools, they do not discriminate against art that is under NDA or covered by copyright.
Computer software company Adobe came under fire recently after it was discovered that users of its product suite had their works automatically opted in to be used for AI dataset development. Art sharing site deviantART also initially opted all published works on the site into AI datasets. Both examples could be undone, but it would require going through a lengthy process to opt out. While deviantART later opted all images out automatically, Adobe has yet to follow suit as of writing.
Additionally, with the development of AI that can synthesize pre-existing voices, it can become much harder to tell fact from fiction.
Voice actor Mike Pollock, best known for his work as Dr. Eggman in the Sonic the Hedgehog series, posted a Tweet calling for YouTube to take down videos that contain faked audio and visuals related to various public figures. Pollock has been openly against using his voice without consent, and several other voice actors share his sentiment, including Steve Blum of Cowboy Bebop fame and Sean Chiplock of The Legend of Zelda: Breath of the Wild fame. Not only can this synthesizer be used to fake audio, but it can also be used to make audio of real people and fictional characters saying offensive terms that the original people never said. Multiple videos created with such tools and malicious intention have already begun spreading around YouTube, most of which feature current US President Joe Biden and former US President Donald Trump making false announcements related to their statuses as presidents.
As AI continues to evolve over time, the boundary between AI-generated art and actual art will shrink, but never to the extent, some speculate it will reach. AI has taken away work from human artists in favor of a more accessible, cheaper way to make art quickly and distribute it to the masses. Even with every advancement in AI, it can never truly replicate human creativity.
Feeding this article into the algorithm would only defeat the entire purpose of its contents.
