Robot Picassos are Taking Over
This technology should bring important ethical problems to our attention
A Muppet DJ made out of mashed potatoes in the style of Van Gogh is now a concept that can be illustrated in sheer seconds with a click of a button, a feat that would take a human artist days or even weeks to complete. Humans grab their paintbrush and let the creative juices flow, but how exactly are these pieces being created by artificial intelligence (AI)?
AI art programs, such as Stable Diffusion and DALL-E 2, use a Deep Learning Algorithm to classify, visualize and organize a large data set of art pieces. This set is created from automated systems, such as LAION-5B that scavenge the Internet for art and feed it to the algorithms. Information such as the artist’s name or subject matter is collected and converted to learning material, informing the program about different artistic aesthetics. This process allows the AI program to convert text to images based on what it has learned.
Famed chatbot ChatGPT’s tremendous success highlights the rise in the popularity and commercialization of AI. However, ethical implications of using AI services, especially AI art, create gray areas as to what can be qualified as original work. This technology is using artists' work without permission or payment to train their models to learn different art forms, which is ultimately comparable to theft.
“Plagiarism isn’t the right word for it,” Brian Patrick Green, the Director of Technology Ethics at the Markkula Center for Applied Ethics, explained. “It is the stealing of the art style of human artists to be recreated, which threatens the livelihood and business of many artists.”
Imitating art styles isn’t confined to machine learning. AI’s ability to replicate aesthetics is far better than what a human copycat can produce, making artists apprehensive to publish their art.
At its core, the issue is that human artists are already undervalued and artificial intelligence seems to only decrease their creative value. Artists must be compensated for their art, especially when technology threatens to take away their artistry–a craft they’ve taken years to hone.
Stability AI, the company behind Stable Diffusion, is currently being sued by three US artists who allege the company “infringed on millions of artist's rights by using their work without permission or compensation to train its model.”
Individual artists aren’t the only ones taking up arms. Getty Images, a multi-million dollar media company, is also suing the company because fragments of Getty’s watermark are present in some outputs–indicating that Getty photos were used to train the algorithm without permission.
Without some sort of payment method or cooperation between these AI art companies and artists, the future of human artwork is threatened, as well as the creation of new art styles. When human artists are uneasy about publishing their own works, they limit the production of new designs. And while these AI art robots can copy other artists’ styles, they cannot form their own.
For regulators of government policy, a difficult balance must be found between the interests of the companies behind AI technology and the artists creating original pieces.
“There are insufficient regulations,” Green explains. “Many governments are trying to figure out what to do with AI and are having trouble keeping up with everything going on from lethal automation systems to AI art.”
Current legislation regarding copyright in the United States stipulates that copyright is only awarded to human-created (not AI), original, tangible and creative works. With no clear indication of ownership of AI-created art, the legal implications of these mechanics are feeble and obscure.
However, this emerging technology is not all negative. We need to find a harmony between artificial intelligence and human use because this technology is revolutionary in how it could change the landscape of the future in our day-to-day lives and industries such as agriculture and healthcare.
“For future technology and human life, AI art has positive broader implications, but we have to make sure not to harm those creating the art in the first place in this process,” Green explains. “As a whole, something that is constant becomes a variable in this big equation that is the world, which brings us into a bigger question of how AI art is going to fit into society.”
AI-generated artwork will be studied in museums and art history classes alongside human-made masterpieces in a decade. But first, we need to take the steps to safeguard human-produced art in preparation for a bigger fight against the gray areas of AI technology.