SoundCloud has addressed growing concerns about the potential use of user-generated music for AI training, confirming that it has not used any artist content to develop AI models, including generative tools.
The platform updated its terms of service in February 2024 to clarify how user-uploaded content might interact with AI technologies within SoundCloud itself, but the company insists that it has not yet used music for AI training.
Marni Greenberg, SoundCloud’s SVP and head of communications, explained that the company has implemented safeguards, including a “no AI” tag, to prevent unauthorized use of content for AI training purposes.
SoundCloud does use AI for certain functions, such as personalized recommendations and fraud detection, but assures users that it has no plans to use music for generative AI development at this time.
While the terms of service now allow the possibility of AI-related use for certain types of content, they explicitly prohibit the use of licensed music for AI training, including for generative AI models.
In the future, if SoundCloud decides to explore AI training using user content, it promises to introduce opt-out mechanisms for users and remain transparent with the creator community.
However, some critics have pointed out that SoundCloud’s communication regarding the terms update was not as transparent as it could have been.
Tech ethicist Ed Newton-Rex highlighted that there were no direct emails alerting users about the changes, even though the platform claims it will provide “prominent notice” for significant terms changes.
Despite the current clarity, SoundCloud’s future actions in this area will be closely watched by its user base, especially as the potential for generative AI continues to grow.
