The music industry faces an unprecedented standoff as record labels strike deals with artificial intelligence companies to develop new generative music tools, while the artists whose performances underpin these technologies increasingly refuse participation. Universal Music Group, Sony Music and Warner Music Group control vast catalogues of commercial recordings, granting them certain licensing privileges, yet the singers and musicians behind these works remain unconvinced that AI represents a worthwhile opportunity rather than an existential threat to their careers and creative control.
The tension reflects a fundamental misalignment of interests. While record labels have grown eager to demonstrate AI competence to investors—share prices at Universal, Warner and Spotify Technology have tumbled amid uncertainty about the sector—musicians like Madonna and SZA have made their positions unmistakably clear. Madonna's manager Guy Oseary declared on the Tim Ferriss podcast that the superstar categorically rejects any AI training of her music, regardless of financial inducement. SZA took to Instagram with a blunt message after discovering her work appeared in training datasets exposed by The Atlantic: "There's nothing you could ever say to me to make this okay." These aren't peripheral figures; they represent a significant swath of the music industry's most commercially valuable talent.
The legal and financial frameworks governing AI's use of music remain unsettled. Artists and their representatives harbour legitimate concerns about compensation models and protective mechanisms. Many performers are willing to explore AI opportunities but only after establishing clear terms that guarantee payment for derivative works and robust safeguards around voice and likeness rights. This deliberate caution contrasts sharply with the speed at which executives are announcing partnerships. Labels have inked agreements with platforms like Udio and Suno without publicly disclosing artist participation, creating an accountability vacuum that fuels suspicion among the creative community.
Record labels have attempted to navigate this minefield through selective disclosure and reassurance. Michael Nash, Universal's chief digital officer, claimed in July conversations with thousands of artists and their estates had yielded many willing participants, though he offered no verifiable details. Robert Kyncl of Warner acknowledged the process remains "complex and laborious" but insisted his company is working toward securing proper permissions. These statements ring hollow to many artists who feel overlooked in the rush to commercialise their work. The industry's habit of announcing major technological partnerships before addressing artist concerns has bred distrust and appears to confirm fears that corporate interests supersede creative rights.
The distinction between training AI models and generating new content deepens the conflict. Record labels argue they possess authority to license existing recordings for algorithmic training without explicit artist consent, though ethical concerns have prompted some to seek permission regardless. However, the more contentious application involves enabling users to generate entirely new compositions in the style or voice of specific artists—essentially allowing someone to type "write a song about a beach day in Taylor Swift's voice." Artists recoil from this prospect because their voices represent irreplaceable, singular assets. Voice cloning technology raises particular alarm, as musicians fear losing control over what statements, messages or artistic directions might be attributed to them through synthetic performances.
The litigation landscape complicates negotiations further. Prior to announcing commercial partnerships, both Warner and Universal sued AI startups including Udio and Suno for alleged copyright infringement. These legal actions continued even as the same labels negotiated licensing deals with the defendants, reflecting the contradictory positions companies must adopt. Sony Music has pursued a more cautious strategy, maintaining active litigation against both platforms while limiting its commercial entanglements. This fractured approach leaves artists uncertain about whether label agreements protect their interests or simply shield corporate partners from legal exposure.
The Atlantic's June publication of standard training datasets illuminated a previously obscured reality for the artist community. The searchable database revealed which musicians' catalogues had been incorporated into major AI systems without their knowledge or consent. This transparency galvanised opposition and forced conversations that labels might have preferred to keep confidential. Artists discovered they had already been used to build commercial technology, sharpening their determination to control future applications of their work.
For Southeast Asian readers and particularly Malaysian music professionals, these developments carry significant implications. As the region's music and entertainment industries expand, questions about how local artists will be protected under emerging AI frameworks become increasingly urgent. If international standards for artist compensation and consent remain underdeveloped, Malaysian and regional musicians risk seeing their work incorporated into global AI systems with minimal benefit. The precedent being set in Western markets will likely influence how technology companies approach licensing in developing economies, where regulatory oversight may be even weaker.
The standoff suggests the music industry requires new institutional structures. Artists need enforceable mechanisms to consent to specific uses of their work, transparent accounting of how their performances contribute to AI system training, and compensation formulas reflecting the actual value those performances generate. Technology companies need legal certainty about what they can use without risking litigation. Record labels need to reconcile their licensing authority with evolving artist expectations around creative control. Governments and international bodies will eventually need to establish baseline standards.
Until such frameworks materialise, expect continued friction between labels announcing AI capabilities and artists withholding permission. The current moment reveals that raw corporate power—the ability to control distribution and exploit existing contracts—cannot substitute for genuine artist collaboration in legitimising new technologies. Madonna's categorical rejection and SZA's public condemnation suggest that major artists possess sufficient leverage to force meaningful negotiation rather than accept whatever terms emerge from boardrooms. Smaller artists lack such bargaining power, raising equity concerns that will likely become more acute as AI deployment accelerates across the industry.
