Music labels and tech companies are addressing the unauthorized use of artists' work in training AI music generators like Udio and Suno. Recent settlements with major labels aim to create new revenue streams, while innovative tools promise to remove unlicensed content from AI models. Artists remain cautious about the technology's impact on their livelihoods.
The rise of AI-generated music has sparked concerns over copyright infringement, as companies such as Udio and Suno have trained their models on billions of data points from human-made songs without artists' permission. These AI systems, including others like Soundful, Boomy, Musicfy, and Playbeat, analyze song structures statistically to produce new tracks based on user prompts, such as creating an anthemic song in the style of The Strokes, The White Stripes, and Arcade Fire. Critics argue this process amounts to sophisticated regurgitation rather than true innovation, with no risk of AI producing figures like Paul McCartney or Bob Dylan.
Legal challenges have emerged, but progress is underway. Universal Music Group and Warner Music, two of the three major labels, recently settled with Udio, promising new licensing agreements to open revenue streams for artists in 2026. Sony Music, however, continues its litigation. Proponents suggest artists can opt out of training data, though labels warn this might mean forgoing compensation. Canadian singer-songwriter Mac DeMarco expressed skepticism, telling The Globe and Mail, “Soon we’ll all just be batteries, like in The Matrix.”
Artists worry about undetected scraping of their work and its potential to undermine their control over intellectual property, echoing unfulfilled promises from the streaming era. Solutions are emerging: Musical AI, backed by new funding, offers an attribution model to identify influences in AI outputs and enable takedowns. Meanwhile, Israeli firm Hirundo develops 'machine unlearning' technology, likened by co-founder Ben Luria to the neuralyzer from Men in Black. Luria explained, “As generative AI systems become more capable, they are increasingly colliding with copyright law — especially in music and other creative fields.” He added, “Creators should be able to say, ‘I didn’t consent to this,’ and there should be a practical way for companies to actually fix it — not just apologize for it.”
This situation parallels the 1980s sampling disputes, where technology outpaced regulation, leading to new protocols. Experts anticipate a similar resolution for AI, potentially taking years to stabilize the industry.
