Meta has released an AI-powered music generator called MusicGen, which can turn a text description into about 12 seconds of audio. The tool can be steered with reference audio, in which case it’ll try to follow both the description and melody. MusicGen was trained on 20,000 hours of music, including 10,000 high-quality licensed music tracks and 390,000 instrument-only tracks from Shutterstock and Pond5. The company has made available pre-trained models that anyone with the right hardware can run. MusicGen’s songs are reasonably melodic, at least for basic prompts, and on par with the results from Google’s AI music generator, MusicLM. MusicGen surprisingly outshined MusicLM in terms of musical coherence when given a more complicated prompt. However, generative music raises major ethical and legal issues, as AI like MusicGen learns from existing music to produce similar effects. There is still a lack of clarity on whether deepfake music violates the copyright of artists, labels, and other rights holders. Several lawsuits making their way through the courts will likely have a bearing on music-generating AI. Meta is not imposing restrictions on how MusicGen can be used.
