The ability of Stable Audio by Stability AI has been further improved after Stable Audio 2.0 was launched a while ago. This music generative AI can be 3 minutes of music and it's also free to use.
Music can be generated using demos or music samples that are not copyrighted. At this point music can be generated but without any lyrics. According to Stability AI, Stable Audio was trained using 800,000 pieces of music from AudioSparx. Only non-copyrighted music is used to train the model to ensure there are no future lawsuit issues.
Last year Stable Audio 1.0 was launched but with a limit of generating music of 90 seconds. Since then several more generative AI into music have been launched to allow users to create copyright-free music for personal use.