<\/p>\n
Universal Music Group and Warner Music Group are reportedly in talks to secure what could become the music industry\u2019s first major AI licensing deal. According to the Financial Times<\/a>, a \u2018landmark AI licensing deal\u2019 could be just weeks away. That\u2019s an agreement that may completely change how music and AI interact.<\/p>\n The goal? To prevent the unlicensed use of copyrighted music by AI companies when training their models. Just take a look at the major\u2019s lawsuit against Suno, one of the biggest AI music generators, who recently amended their lawsuit<\/a> to include stream ripping allegations.<\/p>\n Instead, a licensing deal looks to ensure artists and rights holders stay in control and get compensated when their work influences AI-generated music.\u00a0<\/p>\n Discussions are already underway between the labels and several major tech companies and AI startups, including ElevenLabs, Google, Spotify, Stability AI, Suno, Udio, and Klay Vision.\u00a0<\/p>\n The report also indicated that Sony Music is reportedly in talks with companies that have \u201cethically trained models\u201d, though the label is yet to commit to anything concrete.<\/p>\n Details are still under wraps, but the broad idea is clear: labels want to create an AI licensing model similar to streaming. That means AI companies would need to pay for the use of copyrighted songs when training their models. It doesn\u2019t stop there though. It would also involve micropayments to the tracks used to influence an AI-generated song, each time it\u2019s played.\u00a0<\/p>\n Essentially, AI-generated music is influenced by human-made songs, and the artist behind that original work could get paid.<\/p>\n However, this isn\u2019t an easy task. That would require technology capable of identifying which songs were used to influence each AI-generated piece.<\/p>\n Just last week, Universal Music Group and Sony Music announced a partnership with SoundPatrol<\/a>, a company developing technology to detect the influence of copyrighted works in AI-generated tracks. This kind of attribution tool could make compensation possible, and pave the way for licensing models to function effectively.<\/p>\n While the technology isn\u2019t live yet, it represents a crucial step toward potentially protecting artists\u2019 work in the age of AI.<\/p>\n AI-generated music is flooding streaming platforms. Deezer<\/a> recently revealed that nearly 30% of new tracks uploaded daily are AI-generated<\/a>, and an AI artist has even secured a multimillion record deal. So far, legal systems and streaming platforms have been slow to react, leaving artists vulnerable to having their work used without consent or compensation.<\/p>\n If major labels can agree on a licensing approach, it could finally bring structure to this new creative landscape. For artists and labels, that means more control, more transparency, and compensation for the influence their work has on AI models.<\/p>\nWho\u2019s involved?<\/h3>\n
What could these deals look like?<\/h3>\n
The tech that could make it possible<\/h3>\n
The bigger picture<\/h3>\n
\n