SoundCloud won't train AI on uploads without consent

The tool moment

SoundCloud states, in its own current legal terms, that it will not use what artists upload to train generative AI without their consent. The current Terms of Use, last amended in August 2026, state: 'We will not use your content to train generative AI models that aim to replicate or synthesize your voice, music, or likeness without your explicit consent, which must be affirmatively provided through an opt-in mechanism.' The same section states that neither SoundCloud nor any third party may use content delivered under separate agreements with rights holders 'for the purposes of informing, training developing (or as input to) artificial intelligence technologies without authorization from the applicable rightsholders.' A dedicated page, SoundCloud's AI Partners page, restates the commitment more plainly: 'SoundCloud will not train generative AI models on user content without explicit consent. Period.'

What the documents show

This is a live policy commitment, not a fixed one-time announcement. SoundCloud's own archive of past Terms of Use versions shows no opt-in training clause in the October 2022 version, only a narrower rule against using generative AI to impersonate a voice or likeness by February 2024, and the specific opt-in language present from the May 2025 version onward. The wording a reader sees today reflects a policy revised more than once, most recently in August 2026, rather than a single fixed earlier statement; this note describes the commitment as SoundCloud's current materials state it, retrieved 16 September 2026, rather than dating it to an earlier announcement this note could not confirm from SoundCloud's own pages.

What stays with the musician

A platform-level promise not to train on uploads without consent covers what SoundCloud itself does on its platform; it does not extend to a separate production tool an artist uses before uploading, and does not retroactively address content a different service may have already used. An artist relying on this commitment still needs to check that any opt-in setting SoundCloud provides is configured as they intend, and should not assume it applies to a DAW plugin, a mastering service, or another platform's own terms.

Judge it by listening

There is nothing here for a listener to evaluate by ear; the commitment is contractual, not sonic. Editorially, the practical test is not listening but checking: does the opt-in setting on their own account reflect their intent, and have they read the current terms directly, rather than a summary of what SoundCloud is reported to have said, since the version history shows this language changing over time.

• Does my SoundCloud account's consent setting for AI training actually reflect what I intend, rather than a platform default?
• Does this commitment cover the platform I uploaded to, or does a tool used earlier in production have its own, different policy?
• Have I read the current Terms of Use directly, given that SoundCloud's own version history shows this language has changed more than once?

SoundCloud's opt-in commitment is a platform-level provenance rule stated in its own current materials, distinct from, and no substitute for, whatever a production tool used earlier in the process states about its own training data.