The clash between the music industry's drive to capitalise on artificial intelligence and artists' refusal to surrender their creative work without compensation has reached a critical juncture. Record labels including Universal Music Group, Sony Music and Warner Music Group are racing to license their vast catalogues to AI companies, yet the performers whose artistry built these collections have not consented to the arrangement. This fundamental disconnect between corporate strategy and artist autonomy is reshaping conversations about intellectual property, consent and fair compensation in an increasingly automated creative landscape.
Madonna's stance encapsulates the artist resistance most plainly. Through her manager Guy Oseary, the pop icon declared on the Tim Ferriss podcast that no financial incentive would persuade her to participate: "Don't care what you want to pay her. She's very clear. I do not want my music to be trained on. I want my music to be its own thing." Her position is not idiosyncratic. Prominent musicians like R&B artist SZA have voiced equally emphatic opposition, with SZA writing on Instagram that nothing could justify the practice after discovering her work had already been incorporated into AI training datasets without her knowledge or approval.
The technical and legal architecture underlying these disputes reveals why artists remain so wary. While Universal, Warner and Merlin have already signed agreements with AI platforms Udio and Suno Inc—services that generate original songs from text prompts—these arrangements did not require artist participation. The record labels assert they possess the rights to license their catalogues, but artists and their representatives argue that voice cloning and synthetic performances constitute a separate category of intellectual property that demands individual consent and compensation frameworks.
What makes the situation particularly contentious is that neither clear financial models nor protective legal structures exist yet. Artists considering AI collaboration want guarantees before proceeding. They seek assurance that they will receive royalties whenever their voice or likeness appears in generated content, and they demand mechanisms to prevent unauthorised use of their identity in contexts they never approved. Until such frameworks materialise, most artists view AI partnerships as existential threats rather than opportunities, particularly because their voices are irreplaceable and singular assets unlike traditional songwriting or production credits.
The record labels' own conduct reveals their discomfort with the current situation. Universal Music Group and Warner Music Group initially sued both Udio and Suno for copyright infringement before subsequently negotiating licensing deals with them. Sony Music adopted a more cautious approach, continuing litigation while exploring selective partnerships. These legal reversals suggest that the companies recognise the fundamental unfairness of their position and are now seeking retroactive legitimacy through commercial arrangements. Yet executives have announced these AI agreements whilst remaining evasive about artist participation. Universal's chief digital officer Michael Nash claimed the company had secured commitments from "thousands of our artists and their estates," but declined to name a single participant.
Stock market reactions underscore investor anxiety about these unresolved disputes. Universal, Warner and Spotify have all experienced significant share price declines as market participants weigh the reputational and legal risks of proceeding without artist consent. This financial pressure creates incentives for all parties to negotiate, yet also tempts corporations to move forward regardless, banking that retrospective settlements will prove cheaper than halting development.
The distinction between training AI models and enabling user-generated content based on artist identities matters considerably. Whilst training datasets raise copyright questions primarily, allowing users to generate music "in the voice of Taylor Swift" or similar artists introduces personality rights and defamation concerns. Artists particularly resist this second application because synthetic voice technology creates unprecedented capacity for impersonation and for putting words and sentiments into their mouths they never endorsed. The reputational risks compound financial ones.
For Malaysia and Southeast Asia, these developments carry particular significance. The region's creative industries are developing rapidly, with growing numbers of local musicians, producers and sound engineers building international audiences through digital platforms. If multinational AI companies establish training regimes without artist consent in Western markets, the precedent will likely extend to the Global South. Malaysian and regional artists could find their work incorporated into AI training datasets with even less negotiation or compensation than their Western counterparts experience. The absence of robust artist protections now will leave Asian creators especially vulnerable.
Moreover, the broader implications for creative labour standards in Southeast Asia are troubling. If AI companies can access and process catalogues without permissions, the economic value of artistry itself diminishes. Musicians already struggle with streaming platforms that pay fractional royalties; AI training without compensation would accelerate the erosion of sustainable income for professional musicians. Studios and labels in developing economies lack the bargaining power of Western corporations, meaning local talent could be exploited systematically.
The record labels' announcement of Spotify's AI remix feature, developed alongside Universal and Merlin, demonstrates how quickly these applications proliferate once training occurs. Users will soon be able to modify and redistribute existing recordings through generative AI, creating new versions without artist involvement. This capability further motivates artists to resist participation altogether rather than negotiate piecemeal agreements with individual platforms.
The path forward remains uncertain. Stakeholders increasingly acknowledge that sustainable AI integration in music requires establishing industry standards, collective licensing frameworks, and transparent royalty distribution mechanisms before deployment accelerates further. Industry bodies, artist representatives and technology companies are negotiating these structures, but progress remains slow. Until comprehensive agreements materialise, expect continued resistance from established artists and heightened litigation over copyright and personality rights. The outcome will establish precedents affecting creative workers globally for decades to come.
