When the Muse Becomes the Machine: How AI Is Changing the Rules of Creativity
Multiverse. (Photo/Pixabay)
By Kankon Sen
With over a million streams, two albums on Spotify and a feature in Rolling Stone, psych-rock band The Velvet Sundown seemed like the breakout band of the summer. The catch? The band didn’t really “exist.” It would only be revealed after months of denial that it was a “synthetic music project guided by human creative direction, but composed, voiced and visualized with AI support”. In other words, the hottest new act wasn’t actually human.
In the age of algorithms, an AI-generated band shouldn’t come as a surprise. Nevertheless, the rise of AI invites a series of questions around the new rules for creativity: When an AI model generates a work of art, who gets the credit? Can creators restrict their art from being used as training data for algorithms? And what does AI imply for the future of human-created art?
This friction between artists and AI on copyright issues is now quite palpable. Recent flashpoints have ranged from 6,500 creatives objecting to Christie’s AI Art Auction for using their copyrighted material to train GenAI models, to Disney and Universal suing an AI-image generating firm for reproducing its characters and visual styles without a license. The common thread across these tensions is quite simple: the creative economy has limited clarity of ownership over artwork.
Without clarifying ownership, creators struggle to protect, license or monetize their work. International policy has yet to catch up. Regulation is still scattered across countries’ competing priorities for copyrighting AI-generated art. While Europe focuses on transparency, countries like Japan and Singapore have chosen to allow AI training on copyrighted data under certain conditions. Yet, many jurisdictions, like Mexico and the Czech Republic, still go with the default option: only humans can hold copyright ownership. As a result, fully AI‐generated works often go unprotected despite their monetary value. Until there’s coherent, international regulation in the global creative economy, companies, regulators, and the public remain in limbo.
So where does this haze of regulatory uncertainty leave artists? Generative AI companies rely upon millions of data points from musicians, illustrators, artists or writers to build the nifty text-to-image tools (like Midjourney), and simplified video generators (like Sora) that millions of users use every day. Developers view this from a “fair-use” perspective, which argues that most training data is publicly accessible and crucial for inspiring real innovation. Moreover, the sheer volume of training data typically ensures that the final product looks nothing like its many, many muses; many fair-use debates have been settled with this viewpoint.
Still, artists should be able to draw their own boundaries on folding their creations into these algorithms, because some trends do mimic unique styles closely. For instance, the recent Studio Ghibli wave saw everyone turning themselves into dreamy, pastel-colored “Ghibli-style” characters for profile photos, personal brands, and even LinkedIn headshots. But the trend also stirred some commentary: was it a fun tribute, or a case of misusing artists’ work? The “fair use” argument is far from settled.
Finally, to quote ABBA: it would be remiss to ignore the “money, money, money” of it all. Research from the International Confederation of Societies of Authors and Composers (CISAC), representing over 5 million creators, predicts that the generative AI market for music and audiovisual content will grow from its current value of €3 billion to a whopping €64 billion by 2028. On the flip side, the same research also suggests that music creators are at the risk of losing 24% of their income by 2028.
This tension strikes at the heart of a much bigger question of how we value human creativity. When AI-generated music and art floods our screens, audiences and algorithms risk prioritizing the amount of content created over authenticity. Granting free reign to AI content generators encourages faster innovation and translates into subscriptions, API sales, and reduced operational costs for many stakeholders. But without regulations, artists will have fewer opportunities and performers, producers, authors, and songwriters will increasingly struggle for fair compensation.
Academics (and policymakers) are earnestly exploring new licensing frameworks like “AI royalties” and similar compensation schemes to pay creators when their work is used to train models. Globally, policymakers, lawyers, creatives and AI companies are also navigating the grey areas of ownership and fair use. The next few years are crucial to ensuring that innovation doesn’t come at the cost of the people who make creativity possible. By the time The Velvet Sundown plays its encore, we hope to have better regulations that encourage innovation while preserving artistic integrity.
1. These generators allow you to input a script, blog post, or topic, and the AI tool will generate a complete video, including visuals, music, and voiceovers.
2. The "Studio Ghibli wave" refers to a recent viral trend where users generated images in the distinctive style of Studio Ghibli using AI. This trend gained massive popularity after OpenAI CEO Sam Altman shared a Ghibli-style image of himself. Studio Ghibli, Inc. is a Japanese animation studio founded by directors Hayao Miyazaki and Isao Takahata and producer Toshio Suzuki in 1985.