The music industry faces mounting tension as record labels race to monetise artificial intelligence opportunities while artists increasingly resist having their work used to train the technology. Universal Music Group, Sony Music and Warner Music Group control vast catalogues that they can theoretically license to technology companies, yet a critical gap remains: the performers whose voices and artistic identity gave those recordings value have not agreed to participate. This fundamental disconnect between corporate licensing power and individual artist consent is reshaping negotiations across the industry and raising questions about creative control in the digital age.
Artists ranging from global superstars to emerging performers have made their opposition clear. Madonna's representatives stated unequivocally that she will not allow her music to be used for AI training regardless of financial incentive. Singer SZA went further, publicly condemning the practice after discovering her work had been included in training datasets without her knowledge. Their resistance reflects broader concerns within the creative community about surrendering control over their most valuable asset: their distinctive voice and artistic identity. Many musicians worry that permitting AI systems to replicate their vocal characteristics could enable unauthorised content generation that damages their brand and creative integrity.
The hesitation extends beyond straightforward objections to the technology itself. Even artists willing to explore AI partnerships are proceeding cautiously, recognising that the legal and financial frameworks governing such arrangements remain fundamentally underdeveloped. Musicians and their representatives want clear guarantees about compensation whenever AI-generated content using their voice is monetised. They also seek enforceable protections for their name, image and likeness, ensuring they retain control over how their artistic identity is deployed. Until these safeguards are formalised, many creators view participation as an unacceptable risk with unknown long-term consequences.
Meanwhile, the record labels themselves have navigated a peculiar position, announcing AI partnerships with startups like Udio and Suno without first securing explicit artist commitments. Senior executives have suggested that artists broadly support these initiatives, yet they have declined to name any participating performers. Michael Nash, Universal's chief digital officer, stated in July that the label had engaged in extended conversations with thousands of artists and their estates, claiming substantial buy-in. Warner's chief executive similarly framed the company as working to streamline permission processes. These assurances appear designed primarily to reassure investors rather than address creative community concerns, highlighting the tension between shareholder expectations and artist interests.
The commercial stakes driving this urgency are substantial. Streaming services, record labels and technology firms have invested heavily in AI music capabilities, and investor confidence depends on demonstrating viable business models. Concerns about the sector's trajectory have already triggered significant stock price declines for Universal, Warner and Spotify, creating pressure to announce partnerships and monetisation strategies. This financial imperative has led companies to prioritise corporate licensing agreements over the more tedious work of securing individual artist permissions, establishing compensation structures and developing legal protections.
Interestingly, the record labels themselves initiated legal action against some AI music platforms before pivoting to licensing deals. Universal and Warner previously sued Udio and Suno for copyright infringement, arguing that the companies trained their models on protected works without authorisation. Sony Music has maintained a more measured approach, continuing active litigation while exploring selective partnerships. These shifting positions suggest that corporate strategy remains unsettled, with labels simultaneously treating AI startups as threats and business partners depending on circumstances and negotiations.
The deals announced thus far involve both music generation platforms and remix features. Udio, which enables users to create songs through text prompts, has secured agreements with Warner, Universal and Merlin, the organisation representing independent labels and distributors. Suno, offering similar generative capabilities plus download and sharing functions, reached separate arrangements with Warner. Spotify is collaborating with Universal and Merlin to develop an AI remix feature. However, these partnership announcements obscure an uncomfortable reality: the labels have not disclosed whether individual artists have actually authorised their participation, raising questions about the legitimacy of claims regarding artist consent.
A particularly revealing moment came when The Atlantic published in June a searchable database of training datasets used by popular AI music models. The publication exposed which artists' works had been incorporated into AI systems without clear consent or compensation frameworks. The database demonstrated the scale at which music had been harvested for training purposes, alarming performers who discovered their work included without their knowledge. SZA's response was blunt, posting on social media that no explanation would justify the practice. Her reaction captured the sentiment of many musicians who felt their creative rights had been violated in service of corporate profit motives.
The controversy intensifies when considering what AI companies ultimately seek to accomplish. Beyond simply training algorithms on existing music, the technology firms want to enable users to generate entirely new compositions styled after famous artists. Users could theoretically request songs in Taylor Swift's voice, Madonna's style or SZA's tone, generated by AI rather than performed by the actual artists. This capability poses existential threats to performers' control over their artistic identity and economic interests. A synthesised voice claiming to be an established artist could confuse audiences, dilute the exclusivity of authentic performances and undermine the commercial value of licensing an artist's actual voice for legitimate projects.
Artists worry particularly about synthetic voice generation because vocal distinctiveness represents a core component of their professional identity and market value. Unlike a song composition, which could theoretically exist in multiple versions or arrangements, a performer's voice is singular and irreplaceable. Allowing AI systems to replicate and deploy that voice without restrictions means surrendering control over a fundamental asset. Additionally, musicians fear reputational damage if AI-generated content bearing their synthesised voice delivers messages or appears in contexts they would never endorse. These concerns are not abstract—they reflect realistic scenarios given how rapidly AI voice technology has advanced.
The Malaysian music industry and Southeast Asian artists face particular vulnerability in this emerging landscape. Regional musicians often lack the negotiating power and legal resources of Western counterparts, making them susceptible to having their work incorporated into training datasets without awareness or compensation. As international record labels and technology companies expand AI initiatives throughout Asia, local artists should anticipate similar pressures to license their work or accept the use of their voices in generative AI systems. The current global debate over artist consent and compensation frameworks will establish precedents that directly affect how Southeast Asian musicians can protect their rights and maintain creative control.
Moving forward, the music industry must resolve fundamental questions about artist participation before AI technologies become too entrenched to govern effectively. Creating legitimate frameworks requires transparent agreements about compensation, clear restrictions on voice and likeness usage, and enforceable mechanisms for artist oversight. Without these protections, the current rush to monetise AI could generate lasting damage to artist-label relationships and creative trust across the industry. The coming months will determine whether corporate interests or artist rights ultimately shape how artificial intelligence transforms music creation.
