AI Music Threatens Royalties For Human Artists

The Unsettling Rise of AI on Streaming Platforms

Musicians are sounding the alarm and searching for creative solutions as AI-generated "bands" ascend the streaming charts, consuming a significant portion of already scarce royalty payments. A prominent case is the act known as The Velvet Sundown, which soared to 1.2 million monthly listeners on Spotify, fueling a heated debate about the future of the music business.

As major music unions advocate for legal reform, copyright experts warn that current laws are struggling to address the rapid advancements in artificial intelligence. Allistair Elliott, representing 70,000 musicians in the U.S. and Canada for the American Federation of Musicians, highlighted the core issue: "the technology gets created and gets used before there's any guardrails or policies in place to protect musicians."

Elliott argues that musicians must have the right to consent before their work is used to train AI models and should be compensated if they agree. He stresses the need for a "careful balance" between technological progress and creator protection, fearing that a failure to act could "eradicate" a generation of creative professionals, especially those who rely on recorded music for film, television, and commercials.

Fighting Fire with Fire: A Musician's Creative Stand

Kristian Heironimus, a Florida-based musician performing as Velvet Meadow, found it "disheartening" to see The Velvet Sundown gain a massive following with a name and imagery that felt eerily similar to his own. Instead of succumbing to frustration, he channeled his energy into creating a fuzzy, garage-rock inspired "diss track" titled The Velvet Sundown.

In the song, Heironimus sings, "Pick a real guitar to play, not a damn mouse. Cut your teeth just like me, and all the artists whose fingers bleed." By doing this, he is cleverly using the AI band's name to boost his own song in the algorithm, effectively challenging the machine at its own game. "As artists, we just have to outperform [AI] the best we can," he said, expressing optimism that this challenge will spur a new wave of uniquely human sounds.

The Velvet Sundown uses this AI-generated image as its profile photo on Spotify and social media. (Facebook)

The Financial Squeeze and Global Spread of AI Music

The trend of AI-generated music is rapidly going mainstream. Following the Velvet Sundown controversy, India's Collective Media Network launched its own AI "spiritual" rock band, Trilok. Meanwhile, the streaming platform Deezer reported in April that AI-generated tracks account for 18 percent of its daily uploads. Other major platforms like Spotify and Apple do not currently label AI-generated content.

New York musician and vlogger Hadji Gaviota, who explored the issue in his video Spotify's Fake AI Band Problem, stated that while artists have long known streaming isn't a primary income source, the influx of AI acts makes the situation worse. Because Spotify's royalties are paid from a shared revenue pool, more AI music means "an even smaller slice of the pie" for human artists.

Hadji Gaviota says the deluge of AI music makes it clearer that artists can't rely on streaming income. (Submitted by Hadji Gaviota)

The crisis has led to a wave of lawsuits. Major record labels like Universal, Warner, and Sony are suing AI music tools Suno and Udio for mass copyright infringement, accusing them of training models on their recordings without consent. However, Gaviota is skeptical, noting that the labels are also seeking equity in these AI companies as part of the legal action. "Artists getting a fair deal is usually like the last part of anything," he commented.

Navigating a Legal Labyrinth in the AI Era

Mira Sundara Rajan, a copyright expert and visiting law professor, says that conversations around regulating AI have been "going in circles" for years. She notes a reluctance in the legal community to criticize AI for fear of being labeled anti-technology. While intellectual property laws exist, they are difficult to apply to AI. "You don't have catch-all legal provisions," she explained.

Intellectual property lawyer and classical pianist Mira Sundara Rajan notes the difficulty of stretching existing laws to cover AI. (Submitted by Mira Sundara Rajan)

There is ongoing debate about whether scraping artists' music to train AI models constitutes "fair use" in the U.S. or "fair dealing" in Canada. Sundara Rajan argues firmly that it qualifies as neither. "On any reasonable understanding of copyright law as it stands today, this is not fair use... and it's definitely not fair dealing in Canada," she stated.

She points out that large tech companies wield significant lobbying power, often shaping regulations in their favor, with Canada and other countries frequently following the U.S. lead. This creates a "vicious-cycle situation" where the two groups that lose out are the artists and the public, who lack the power to influence policy.