SoundCloud has issued a statement clarifying its position on the use of user-uploaded music for training artificial intelligence (AI) models. The company maintains that it has never used user content for this purpose and currently employs technical safeguards to prevent unauthorized access. This follows a report by TechCrunch highlighting a February 2024 update to SoundCloud’s terms of service, which seemingly opened the door for future AI training on user content.
However, SoundCloud SVP and head of communications, Marni Greenberg, assures users that the terms of service update was intended only to clarify how content might interact with AI technologies *within* the SoundCloud platform itself, for uses such as personalized recommendations and fraud detection. Greenberg explicitly stated that SoundCloud does not develop AI tools and actively prohibits third-party scraping of its content for AI training purposes.
The company’s stance on using user-uploaded music for generative AI development is nuanced. While the terms of service explicitly forbid the use of licensed content (music from major labels) for AI training, they do allow for the possibility of using other types of uploaded content. Importantly, SoundCloud emphasizes that no such use has occurred to date. Greenberg further pledged that if SoundCloud were to ever consider using user content for generative AI training, robust internal permissioning controls and clear opt-out mechanisms would be implemented in advance.
This commitment to transparency is crucial, especially considering that the previous terms of service update apparently went largely unnoticed by many users. Tech ethicist Ed Newton-Rex, who initially brought the terms of service changes to light, reported receiving no email notification about the alterations, a sentiment echoed by others. While SoundCloud claims to provide ‘prominent notice’ of significant terms of service changes, the lack of widespread awareness raises concerns about the effectiveness of their notification process.
The situation highlights the ongoing tension between the potential benefits of AI and the need to protect user data and intellectual property. SoundCloud’s assurances are welcome, but the lack of previous transparency and the potential for future AI training on user content warrant continued vigilance and a demand for clear, proactive communication from the platform. Users should stay informed and actively monitor SoundCloud’s updates to ensure their rights are protected.
