Arooj Aftab Calls for Artist Pay When Music Trains AI Models

Grammy-winning Pakistani singer Arooj Aftab has called for musicians to be compensated when their recordings are used to train artificial-intelligence systems. Her intervention adds a prominent Pakistani voice to an international dispute over copyright, transparency and the distribution of streaming revenue as AI-generated music becomes increasingly common.

Aftab raised the issue through two Instagram Stories posted on Tuesday. She shared material from United Musicians and Allied Workers, a trade union representing music-industry professionals, which demanded clear labelling of AI-generated music throughout its distribution so listeners and platforms can distinguish it from work created by human artists.

The singer argued that generative systems depend on existing human-made recordings to produce new music. In her view, using an artist’s catalogue as training material constitutes a form of commercial usage and should create an obligation to compensate the musicians whose work helped the technology learn musical patterns, styles and structures.

Aftab also supported the union’s call for streaming platforms to direct payments to human artists rather than creators of fully AI-generated tracks. She said future regulation must account for the financial impact on musicians and maintained that artists should receive royalties whenever their recordings contribute to derivative works.

The debate has intensified as AI tools become capable of producing songs that imitate familiar voices, genres and production techniques. Musicians and rights groups argue that training models on copyrighted recordings without consent allows technology companies to extract value from creative work while bypassing the contracts and licensing systems that normally govern reuse.

Aftab’s comments are part of a wider international campaign. More than 13,000 artists, including Bjorn Ulvaeus of ABBA and Radiohead singer Thom Yorke, signed a statement in 2024 opposing the unlicensed use of creative works for AI development and warning that the practice threatens artists’ livelihoods.

More than 1,000 musicians, including Kate Bush, later released a silent album to protest proposed changes to British copyright rules that would have permitted AI systems to train on copyrighted music and books. The project used silence as a symbolic warning about the cultural consequences artists fear if their work can be absorbed without permission or payment.

Streaming services have also started introducing measures intended to separate human creators from synthetic content. Spotify launched verification labels for confirmed human artists and separate badges for creators producing AI music, while taking action against low-quality uploads designed to manipulate recommendations and royalty systems.

The platform said it removed more than 50 million spam-like or low-value tracks from recommendations during the year following the policy changes. It also reported verifying over 300,000 human creators, more than 70 per cent of them independent musicians, and said human-made recordings accounted for 99 per cent of user searches.

For Pakistan’s music industry, the dispute raises questions about whether existing copyright protections can address AI training, derivative content and cross-border streaming. Aftab’s position could encourage local artists, rights organisations and policymakers to seek clearer consent, labelling and payment rules before generative music becomes more deeply embedded in commercial platforms.