What Artists Fear More Than a Blank Canvas: AI
Why do artists fear artificial intelligence more than a blank canvas? Because when you confront a blank canvas, it’s just you and your imagination. But with AI? You’re up against a robot that not only matches your every brushstroke but also cheekily suggests, “How about I add a little spaceship in the corner of your Renaissance masterpiece? It could use some intergalactic flair!”
Well, most AI art generators — think Dall-E 3, Midjourney, Stable Diffusion, etc — aren’t that dumb, but the sentiment remains true: most artists don’t believe that these models have any true artistic sense.
AI Training Methods and Artists’ Plight
These AI systems are trained on massive datasets, often incorporating artworks created by various artists across the globe (and time periods), usually without their explicit permission. This practice raises significant ethical and legal questions. For artists, particularly those not widely known, it means their works are being used in a way that they have no control over, potentially affecting their livelihood and undermining the value of their original creations.
Indeed, the legality of using artworks to train AI models without explicit permission remains a contentious issue. Thankfully, in recent months there’s been some clarity, as cases have worked their way through the courts. For instance, on August 18, 2023, the U.S. District Court for the District of Columbia granted summary judgment for the United States Copyright Office in Thaler v. Perlmutter, affirming the Copyright Office’s denial of copyright registration for artwork created by a generative AI system. The court found that the generative AI-created artwork did not satisfy the Copyright Act’s “human authorship” requirements.
That’s at least one part of the gray area illuminated: Based on current law, AI-created artwork cannot be copyrighted, even if a human created the prompt. That provides some relief to artists, in the sense that this task, as of now, cannot yet be automated.
In another significant case, a judge in California federal court trimmed down a lawsuit by artists against Stability AI, Midjourney, and DeviantArt. The artists accused these companies of misusing their copyrighted work in connection with their generative AI systems. While some claims were dismissed, including all allegations against Midjourney and DeviantArt, the key claim that Stability’s alleged use of one artist’s work to train Stable Diffusion infringed her copyrights was allowed to proceed. While the is still ongoing, it illustrates the complex legal challenges in determining copyright infringement in the realm of AI-generated art.
Impact on the Art World
The use of AI in art is not just a legal issue, however, but also an ethical one with significant implications for the art world and artists. When AI models are trained on artworks without proper acknowledgment or compensation to the original creators, it in many ways is a form of exploitation. Artists shouldn’t have their livelihoods and job prospects diminish simply because they share their art with the world for free. Left uncontrolled, this practice will devalue the work of artists, especially those who are emerging or less known, as their unique styles and ideas are co-opted without recognition or financial benefit.
Furthermore, the risk that AI-generated art could flood the market, making it harder for all artists to gain visibility and sell their work, is a growing concern. In addition to the economic impact on artists, this also affects art itself: with fewer artists taking fewer risks, this could lead to a homogenization of art, where the unique, human elements of creativity are overshadowed by algorithmically generated content.
Moral Responsibility
In the rapidly shifting landscape of AI art, it’s crucial to recognize the moral responsibility of AI developers in respecting the rights and creations of artists. This extends beyond legal obligations, as we all know that the law at many times can be inadequate and outdated. Instead, it’s about acknowledging the human creativity and labor that form the bedrock of art itself. Fundamentally, ethical practices in technology development shouldn’t be a fundamental principle in guiding innovation, not an afterthought.
AI developers hold an awesome, profound power now: The ability to shape the future of art and creativity itself. With this power comes the responsibility to ensure that their advancements do not inadvertently harm the very community they trained their models from. This involves recognizing the rights of artists not only in legal terms but also in terms of fairness and respect for their contributions. By doing so, developers can foster a more collaborative and respectful relationship between technology and art, ensuring that advancements in AI art enrich rather than diminish the art world. For example, AI developers should partner with art repositories and websites to create an “opt-in” feature for artists who post their work. These common sense changes, even if in the short-term hurt the bottom-line of these companies, will in the long-run lead to fruitful, productive relationships that pay dividends for these companies and society as a whole.
How We Can Draw Inspiration From Other Fields
A glaring anomaly becomes apparent when we compare the field of AI-generated art with other creative industries like music and literature. In these fields, permission and licensing are standard practices. For instance, a musician can’t use another’s composition without proper licensing, and a writer cannot repurpose another’s text without attribution and permission. These practices are deeply ingrained as they respect the creator’s rights and provide a framework for fair use and compensation.
In contrast, AI art currently operates without these established norms. As I wrote before, artworks are often used to train AI models without explicit permission from or compensation to the original artists. If AI continues to use artworks without appropriate recognition and compensation, it risks creating an environment where artists are undervalued and undercompensated for their work.
The challenge, therefore, lies in bridging this gap. It isn’t an impossible task — far from it. By aligning the practices of AI art creation with those already existing in other creative industries, we can ensure a balanced ecosystem where technology and art coexist synergistically. This alignment would mean:
- Adopting practices like seeking explicit permission
- Providing fair compensation.
- Creating a licensing regime, taking inspiration from other creative fields.
- Acknowledging the contributions of artists in the AI art creation process.
Such measures would not only uphold the ethical standards but also promote a culture of respect and collaboration, essential for the sustainable growth of both art and technology. If we want to preserve the future of artists and creativity, it’s an imperative.
