MAC Joins Other Artist Advocacy Groups in Filing an Amicus Brief in the “In Re Mosaic LLM Litigation”

The Headline: MAC and other advocacy groups have filed a brief arguing against the idea that AI companies can freely use copyrighted music to train their models without artist permission or requisite payment. We are concerned that if courts allow this practice, it will undermine the rights of artists and songwriters to control and profit from their own creative work.

The Music Artist Coalition has filed a joint amicus brief in the United States California Northern District Court, along with the following advocacy groups: American Association of Independent Music, Artist Rights Alliance, Black Music Action Coalition, National Academy of Recording Arts & Sciences, Inc., Recording Industry Association of America, and Songwriters of North America. “In Re Mosaic LLM Litigation” is an important AI related case that caught the attention of music artist advocates for its reliance on the – copying copyrighted works is “fair use” for training AI models – argument.

Like book authors and publishers, amici members represent creators and licensors of compositions and sound recordings who rely on copyright law to keep those works safe from being copied for profit without permission. MAC and others see this as an opportunity to stand up against an attempt to weaken copyright protections, which are essential for all copyright owners.

MAC believes this is an important case for music artists because of the wide-ranging implications that rulings respecting the fair use argument for AI training of copyrighted works can and will have across the industry. Ron Gubitz, Executive Director of MAC said, “We firmly believe artists themselves need to have the opportunity to decide how best to protect the fate of their work. The advancement of AI can be a benefit and a useful tool to music creators and music lovers, but only after it is properly regulated so that artists’ interests, both big and small, are protected.”

The case in question was brought by a handful of authors against Databricks, Inc. and its subsidiary MosaicML. The name of the case is “In Re Mosaic LLM Litigation” in lieu of the traditional Party A vs. Party B because the Northern California Federal District Court combined the various authors' lawsuits into one case for administrative purposes.

The brief we submitted mainly focuses on Databricks, Inc. and MosaicML (the “Defendants”) and provides a counter to their argument that copying for AI “training” is “exceedingly” “transformative” and therefore considered “fair use.” It further examines the alarming market harms that will be caused by the acceptance of widespread unlicensed use of copyrighted works by AI developers such as the Defendants, as well as others working more closely in the music space.

As we know, some labels and publishers have already moved away from litigation and have begun to strike license deals with AI companies. For example, Suno and BMG, Suno and WMG, Udio and UMG, Udio and NMPA. Although MAC previously commented on the UMG/Udio agreement when it was initially announced, the presence of these transactions has only continued, thus if a court were to rule that the training in this case is in fact legal that would seriously threaten this rapidly emerging space in the music industry.

MAC has said from the beginning that as these deals continue the most important thing over all others is that while the executives are cutting deals the artists are not left out of the process. MAC believes any AI music system must enshrine three core principles:

1. Artist Consent: The original music creators need meaningful control over if and how their work is used to train AI systems. 
2. Fair Compensation: Artists must share meaningfully in this revenue, with splits that reflect the fundamental value of their creative work.
3. Deal and data Clarity: Artists need clear visibility into the deals being struck and how their work is being used by the platform.

We will be following the “In Re Mosaic LLM Litigation” and provide updates to our members as necessary.