Universal Music Group (UMG) and Sony Music Entertainment have submitted a legal memorandum requesting the court to unseal the complete inventory of audio files utilized by Suno for training its generative AI models. This action represents a further development in the copyright litigation initiated against the AI music company in 2024.
The total quantity of audio recordings employed in Suno’s model training is designated as the “Model Training Figure.”
The labels’ legal representatives argued in the memorandum that, given the figure’s inclusion in a proposed amended complaint, it should be considered public record, citing legal precedent supporting the strongest presumption of public access for such information.
Suno has asserted that disclosing the figure would result in substantial commercial detriment, citing concerns about competitors replicating its business approach.
Suno has acknowledged using “tens of millions of instances of different kinds of recordings” during the training process.
While UMG and Sony previously redacted the specific number from court filings, they are continuing to challenge the sealing of the information.
The labels have also filed a motion to include an additional 61,026 recordings in the lawsuit, identified through a scan of Suno’s training data using the content recognition service Audible Magic.
