Universal Music Group and Sony Music Entertainment have stated that the public has a right to know the number of recordings utilized in the training of artificial intelligence music models. This position is in opposition to a request from AI company Suno to keep specific training data confidential as part of current legal proceedings.
The record labels, along with other plaintiffs, have filed court documents arguing against Suno’s redaction of figures detailing the scale of its training datasets. The case concerns allegations that Suno used copyrighted recordings without authorization to develop its AI music generation systems.
Suno, alongside AI company Udio, has attempted to prevent the release of the number of recordings identified during the discovery phase as having been used in training. The companies maintain that disclosing this information could benefit their competitors by revealing insights into their systems and operations.
Universal Music Group and Sony Music Entertainment countered this argument, referencing Suno’s previous public statements indicating its models were trained on “tens of millions” of works. The labels asserted that a more precise figure would not significantly disadvantage Suno’s competitive standing.
The plaintiffs are also seeking to prevent the redaction of two specific data points: an internal count generated through audio-fingerprinting analysis and the total number of recordings allegedly used in training. The labels explained that the fingerprinting count was derived using Audible Magic technology to identify recordings purportedly incorporated into Suno’s systems.
According to the filing, the fingerprinting-derived number is essential for understanding the extent of the alleged copyright infringement. The labels stated there is “no legitimate basis” for keeping this information from public view.
They also contend that the overall training figure is pertinent to Suno’s legal defense, specifically its claim of fair use. The plaintiffs believe the number will demonstrate the scope of the copying and whether it is limited to their own music catalogues.
Universal Music Group and Sony Music Entertainment further argued that concerns about commercial harm are exaggerated, characterizing the potential impact of disclosure as “at most, marginal.” They also highlighted Suno’s own acknowledgement of using a large volume of data in its training process.
“Having publicly acknowledged copying of that magnitude, Suno cannot credibly claim that disclosing a slightly more precise count of the audio files that its AI model trained on would reveal anything competitively sensitive,” the labels stated in their filing.
This dispute is part of broader legal challenges involving AI-generated music tools and copyright holders, with similar arguments being presented in related cases concerning Musicians’ Union, which shares legal representation with Suno.
A court decision regarding the confidentiality of the figures is pending.
