The music industry has begun banning using its music source data for music-generating artificial intelligence (AI) training. Copyrighted data will change into increasingly difficult to make use of for generative AI training.

On the twelfth (local time), Universal Music Group (UMG), one in all the world’s top three record labels, issued a warning to music streaming corporations equivalent to Spotify and Apple Music to stop using AI-generated music. It was reported that a letter had been sent.

That is the explanation why the music-generating AI may include pop star songs that the corporate owns copyright among the many thousands and thousands of sound source data utilized in the training process. In a letter sent to those corporations, UMG said that they might “not hesitate to take motion to guard the rights of artists” together with these concerns, and demanded that they “block related music downloads.”

Previously, the Recording Industry Association of America (RIAA) identified in an official letter sent to the U.S. Trade Representative that music-generating AI tools produce music very much like the work of famous artists. It was also argued that the AI ​​training process was infringing on copyright by utilizing music without the consent of the association members.

It appears that evidently generative AI is beginning to spread awareness that it is going to not be free from copyright issues since it is trained prematurely with open data on the Web.

This movement first occurred amongst painters. Some painters have filed a category motion lawsuit against the developer, claiming that the image creation tool used their work without their consent. We’re currently awaiting a court decision.

However, ‘Jukebox’, a music generation AI tool released by Open AI, generates a song within the type of the corresponding artist by entering the genre, artist, and lyrics.

Jeong Byeong-il, member jbi@aitimes.com