Google’s new AI-powered text-to-music generator, is now accessible to the public. Called MusicLM, the product is available on the web, Android, and iOS platforms. MusicLM has been highly anticipated as one of the leading generative AI tools for music creation. For producers like myself, I wonder how good it truly is.
Users can easily generate compelling AI-generated music with MusicLM by using simple text prompts such as “melodic techno” or “relaxing jazz.”
MusicLM can generate music based on a user humming a melody and providing a corresponding text prompt. It can even create music inspired by places, genres, eras, or descriptions of other artworks like paintings.
MusicLM was trained by using 5 million audio clips, this is equivalent to 280,000 hours of listening. Along with this was an additional dataset of 5,500 music clips containing captions written by professional musicians. Each caption provides a description of the music, including its genre, mood, and rhythm.
Potential risks associated with the model include music generation for underrepresented cultures and cultural appropriation. Generative AI has sparked extensive debate within the music industry in recent months. Of course, musicians’ jobs could be in jeopardy.