January 14, 2026
Bandcamp barred substantially AI-generated music from its store
Bandcamp's terms bar training on artists' music without permission, and reported January 2026 statements bar selling substantially AI-made music.
The archive
20 entries shown, each prepared after the events it records and published with its sources attached.
Aiva makes full copyright depend on the subscription tier
Aiva's legal terms make copyright ownership and monetisation of a generated track depend on which paid plan made it.
A character licence, not a contract, built Miku's fan economy
Crypton's own licence and concert pages describe a dual-licence system that has let fans write Hatsune Miku's catalogue since 2007.
September 6, 2023
Deezer and UMG stopped paying noise the same as songs
The companies' joint announcement redefines which streams count as royalty-worthy, and by how much.
September 25, 2025
Spotify set new rules on AI voice clones, spam and disclosure
Spotify's 25 September 2025 announcement documents impersonation, spam-filter and AI-credit policies, without a ban on AI-made music itself.
January 17, 2019
A generation metric adapted from images admits a moderate correlation
The Frechet Audio Distance paper reports a 0.52 correlation with human ratings, calling it better than SDR, not a proxy for listening.
March 16, 2023
Music groups published seven AI principles together
More than 40 organisations signed the Human Artistry Campaign's principles requiring consent and licensing for AI training.
April 12, 2024
A union and the major labels wrote AI consent into a contract
SAG-AFTRA and the major labels reached a 2024 sound-recordings agreement requiring consent and minimum pay before an AI voice replica ships.
November 16, 2023
DeepMind watermarked its own generated music, not music generally
SynthID embeds an inaudible mark in Lyria output; the documents scope its robustness and its limits.
October 29, 2025
A label settled with Udio and announced a licensed platform
UMG's announcement and Udio's own notice describe a 2026 licensed platform, download restrictions, and undisclosed settlement terms.
March 19, 2024
Spotify publishes Loud and Clear as its own royalty evidence
Spotify's Loud and Clear site documents how it defines payouts, the context it supplies, and what a maker should verify before citing its numbers.
January 26, 2023
MusicLM was published with its memorisation study attached
Google's text-to-music paper reported a duplication test and withheld the model for over three months.
September 25, 2025
A DDEX-built credit standard is how Spotify's AI labels arrive
Spotify's own announcement documents an AI-disclosure credit standard developed through DDEX, delivered by distributors rather than typed by a listener.
Melodyne's DNA opened single notes inside a recorded chord
Celemony's own 2008 announcement and 2009 rollout timeline document how Direct Note Access made polyphonic audio editable note by note.
August 5, 2025
Eleven Music named a licensing fee without naming the licensors
ElevenLabs launched a music model with broad commercial claims, then disclosed pass-through fees to unnamed rights holders.
May 3, 2023
Spotify pulled Boomy tracks over suspected stream manipulation
Trade reporting and Spotify's own policy show a 2023 Boomy takedown targeted artificial streaming, not the fact that Boomy's music was AI-generated.
January 24, 2025
Deezer began tagging albums that are fully AI-generated
Deezer's own announcements document a detection tool, a June 2025 tagging system, and the rising delivery share it says the tool measures.
June 11, 2026
A generated stem is not the same object as a session take
Suno’s own description of regenerated stems and Apple’s documentation of session-track export show why the two need different handling.
February 1, 2015
A broadcast standard specifies how to blind-test audio
ITU-R’s MUSHRA recommendation gives a small studio a tested method for comparing generated and human tracks without fooling itself.
July 22, 2024
Open models show two ways to disclose training data
Stable Audio Open and MusicGen name different amounts about their training data, against what the EU’s new code asks providers to describe.