Topic
Text-to-audio models hand you a finished-sounding recording. Text-to-MIDI models hand you notes: pitches, lengths and velocities you can edit, re-voice and play on your own sounds. Here is how the two approaches differ, what research systems such as MusicGen and text2midi actually output, and why the symbolic route keeps more of the song in your hands.
Workflows
Indian law has a clause for computer-generated works, but the first court to look at a Suno-made song was not convinced it is protected at all. Here is how ownership splits between lyrics, composition, recording and voice, what the Delhi High Court actually ordered on 25 Mar 2026, and what to do before you release.
Workflows
Spotify says Artist Profile Protection, now a beta with thousands of artists, will be available to all artists by the end of 2026. We expect it on time. We also make our first scored prediction, about the part Spotify has not promised.
Industry
Labels, publishers and collecting societies have sued Suno, Udio and Anthropic, and independent artists have filed class actions of their own. This page lists 23 major lawsuits, in 21 rows, with their status as of 23 Oct 2026, explains the four legal questions they turn on, and sets out what each outcome would mean for independent musicians.
Workflows
So far, every deal that pays for training AI on music runs through an intermediary (a label, publisher, distributor, collecting society or sample marketplace), and almost none publishes a rate. Here is how the four deal structures work, what the law says in the EU, UK, US and India as of 23 Oct 2026, and what an independent artist can actually claim.
Workflows
One email a week: the music-tech and AI-in-music news that mattered, and what we think about it.