The music industry faces an unprecedented standoff as record labels race to monetise artificial intelligence opportunities while the artists whose creative work underpins these ventures increasingly refuse to cooperate. Over the past twelve months, major labels including Universal Music Group, Sony Music and Warner Music Group have forged partnerships with emerging AI platforms, positioning themselves as technology innovators to reassure investors. Yet this strategic pivot has collided with a fundamental obstacle: the musicians who created the recordings have not authorised their use for training machine learning systems.
The tension reveals a structural flaw in how record labels wield their contractual power. While Universal, Sony and Warner collectively own millions of copyrighted songs and can license catalogues to investors, they cannot unilaterally grant AI companies the right to train algorithms on those recordings without the explicit consent of the recording artists themselves. This distinction between composition rights and artist approval has become the battleground where industry executives and musicians are contesting the future of creative technology. Several prominent performers have made their objections unmistakably clear. Madonna, through her manager Guy Oseary, has declared absolute refusal to participate regardless of financial incentive. The R&B artist SZA has publicly denounced the practice on social media, stating there is nothing companies could say to justify using her voice for AI purposes. Their resistance reflects broader unease within the creative community about surrendering control over personal artistic identity to machines with uncertain commercial and cultural futures.
Those artists willing to explore AI opportunities are nonetheless approaching negotiations with extreme caution. Rather than hastily signing licensing agreements, musicians and their representatives are insisting on establishing clear financial structures and legal protections before committing their voices and likenesses. They want guarantees of payment whenever AI systems generate music using their styles or voices, and robust mechanisms to control how their names and images are deployed. This deliberate stance has frustrated label executives eager to demonstrate AI competency to shareholders. Universal, Warner and the independent distributor collective Merlin have already signed agreements with Udio and Suno, platforms that enable users to generate songs through text prompts. Universal and Merlin are additionally collaborating with Spotify to develop AI remix functionality. Notably, these agreements were announced without the labels simultaneously securing artist commitments, creating a credibility gap between corporate announcements and actual creator participation.
The absence of disclosed artist names among supposedly willing participants raises questions about the substance of label claims. Michael Nash, Universal's chief digital officer, told financial analysts that the company has conducted extensive conversations with thousands of artists and secured opt-ins, yet the company has revealed no identities. Warner Music Group's chief executive Robert Kyncl similarly described negotiations as ongoing and acknowledged the complexity of obtaining permission from individual creators. This opacity suggests that either commitment levels remain lower than publicly stated, or that artists fear professional consequences from refusing deals with their contractual employers. The legal terrain remains equally contested. Labels likely possess the right to license music for training purposes without artist permission, since they hold the copyright to recordings. However, many are voluntarily seeking artist consent because they recognise the reputational risks and ethical concerns surrounding the technology. When The Atlantic published a searchable database in June revealing which artists' works had been used to train popular AI models without disclosure, it exposed the industry's lack of transparency and triggered outrage among affected creators.
The stakes intensify considerably when AI applications move beyond training toward generating content that mimics specific artists' voices and styles. Users want the ability to create new songs explicitly emulating Taylor Swift or other celebrities, specifying vocal characteristics and artistic approaches in their prompts. This capability represents a more profound threat to artists than simple algorithmic training, as it potentially allows third parties to produce content attributed to those performers without involvement or approval. Musicians have been substantially more resistant to granting such voice rights, fearing both loss of creative control and the possibility of their voices being deployed to express views or endorse messages they would never voluntarily associate with themselves. The voice became a particular vulnerability because, unlike visual appearance, one's distinctive vocal qualities cannot be easily altered or reclaimed once distributed digitally.
Stock market movements reveal the broader anxiety permeating the music industry. Share prices for Universal Music Group, Warner Music Group and Spotify Technology have experienced significant declines as investors worry about AI's disruptive potential for traditional music business models. Labels are eager to regain investor confidence by demonstrating early adoption and technological sophistication. Yet they face a paradox: executing AI strategies without artist support undermines the credibility of those very strategies and risks triggering regulatory or contractual challenges down the line. Sony Music has adopted a markedly different approach from its competitors, avoiding sweeping announcements while maintaining active litigation against both Udio and Suno for alleged copyright infringement. This cautious strategy suggests Sony's recognition that the legal landscape remains unsettled and that aggressive positioning could prove costly.
For Malaysian and Southeast Asian audiences, this conflict carries particular relevance as the region has become increasingly important to the global music industry. Local artists, producers and technology companies will inevitably confront similar questions about voice rights, compensation and creative control as AI adoption accelerates across Asia. The precedents being established in negotiations between Western labels and musicians will likely influence how regional disputes unfold. Furthermore, Southeast Asian creators operating independently or through smaller labels may find themselves at a disadvantage if major international platforms implement AI features based on Western negotiated frameworks that don't adequately protect artist interests. The unresolved tension between technological opportunity and creative rights protection will shape whether AI becomes a tool that enriches musicians or one that further concentrates power among platform operators and technology companies.
