A new music-industry system seeks to distinguish synthetic recordings from human-led work

International music organisations have introduced a voluntary labelling framework to distinguish recordings generated substantially by artificial intelligence from those created primarily by humans with limited AI assistance.

The framework introduces two descriptions: “AI-generated” and “AI-assisted.” The Jamaica Observer reports that the first category includes recordings produced entirely through prompts as well as tracks in which lead vocals or key instrumental parts are generated by AI.

The “AI-assisted” designation applies to recordings that remain substantially human-created but contain some expressive elements produced through AI. Under the proposed distinction, the lead vocals and primary instrumental performances would still need to be human-made.

The initiative is supported by major music-industry bodies including IFPI, the Recording Industry Association of America, independent-label organisations, the Recording Academy, SAG-AFTRA and the Human Artistry Campaign. It is intended for broad adoption by streaming platforms and other digital music services.

The labels respond to rapid growth in AI-created music submissions. Deezer has reported that AI-generated recordings account for a substantial proportion of the music uploaded to its service each day. The company already uses detection technology to identify and label synthetic recordings.

The new system focuses mainly on the finished sound recording. It does not fully resolve questions concerning AI-generated lyrics, compositions, artwork, videos or the training data used to build generative systems.

This distinction matters because a recording can be transparently labelled as synthetic without answering whether copyrighted music was used to train the system that created it.

WIPO has argued that generative AI increases the need for effective copyright infrastructure, including rights databases, licensing systems, technical standards and institutions capable of connecting creative use with ownership and compensation.

The proposed labels are useful, but Jamaican creators should not mistake transparency for protection.

A listener deserves to know whether a voice, instrumental or entire recording was generated by a machine. However, the label does not tell the listener which human recordings taught that machine how dancehall drums should move, how a reggae singer phrases a melody or how a Jamaican deejay rides a riddim.

Black music has repeatedly experienced cultural extraction. Rhythms, language, vocal techniques, dances and fashion are adopted globally, while the originating creators receive limited credit or income. AI can reproduce that pattern at industrial speed.

Jamaican music is especially exposed because its influence is far greater than the size of its formal domestic market. Reggae, dub and dancehall have shaped hip-hop, electronic music, reggaeton, jungle, grime, Afrobeats and global remix culture.

A generative system trained on Jamaican recordings could therefore produce commercially useful music while the underlying artists, producers and musicians remain unidentified.

The first line of defence is accurate data. Every Jamaican release should include correct songwriter splits, producer credits, performer information, ISRC codes and documented master ownership. Older catalogues also need digitisation and rights clarification.

Many creators still upload music through distributors without fully understanding metadata. Missing or incorrect information weakens royalty collection and makes it harder to prove that a protected work has been used.

Contracts require equal attention. Artists and songwriters should examine whether label, publisher or distribution agreements include rights related to AI training, synthetic voices, digital likenesses or future technologies. Broad language may authorise uses that were not clearly discussed.

Jamaica also needs a national policy response. JIPO, collective-management organisations, universities and music-industry bodies should provide training on AI rights and develop recommendations suited to the local sector.

The country should participate actively in international discussions rather than waiting for rules designed by major foreign companies. Jamaican music has supplied too much value to global culture for its creators to remain outside negotiations about machine-generated music.

Labels can help audiences make informed choices. Real protection requires consent, licensing, attribution, payment and enforceable rights.