AIs like Chat GPT can do audio, but I think it's mostly in the realm of tracks using an artist's voice singing lyrics to music they've never performed. Meta's AudioCraft AI is designed to create original music using AI trained on content that Meta owns so no (C) issues. It's apparently a couple months old -- it seems Meta's open source announcement coincides with a new version. While there isn't a file you can download to install AudioCraft -- at least not yet -- it's not impossible to use it either. Like most current AIs it likes to use a GPU [graphics processor] rather than a CPU -- a graphics card with at least 16GB RAM is recommended, and Nvidia seems to be favored over AMD -- but you can also run it online using Google's Colaboratory. I've read that using it with graphics cards with less than 16 GB RAM means you'll turn out shorter compositions.
neowin[.]net/news/meta-announces-new-ai-for-realistic-music-and-sound-generation-from-text/
medium[.]com/@alfredolhuissier/how-to-install-audiocraft-meta-ai-music-generation-ai-locally-on-mac-and-windows-too-f340ceb75ada
pypi[.]org/project/audiocraft/
dev[.]to/0xkoji/run-musicgen-on-google-colab-1pjd
colab.research.google[.]com/