Tidal Moves On AI Music With Labels, Fraud Checks And A Royalty Ban

Tidal has published a new policy on AI-generated music, setting out how the streaming service intends to handle a category of content that has grown considerably over the past year.

The policy, dated 29 June, defines AI-generated music as work that is wholly or substantially produced by generative artificial intelligence. Tidal is careful to note that machine learning has long played a part in how records are made. The concern here is narrower, aimed at the recent flood of text-prompted tracks rather than the production tools artists have used for years.

Tidal will continue to accept AI-generated music. The company frames this as a matter of choice, both for artists who want to work with these tools and for listeners who want to decide what they spend their time with. What changes is the level of scrutiny. Tidal says AI-generated content will be held to a higher standard of integrity than the rest of its catalogue.

From mid-July, listeners will see an icon next to anything Tidal identifies as entirely AI-generated. As detection improves, the company intends to extend that label to tracks that are substantially, rather than wholly, machine-made. Tidal is also clear that it does not want to carry this responsibility on its own. It expects distributors to flag AI content before it reaches the platform, and says it will begin enforcing that expectation.

The firmest line concerns fraud. Tidal will block or remove AI-generated music tied to deceptive behaviour, naming impersonation, attempts to interfere with genuine artists and their audiences, and the familiar pattern of bulk uploads paired with suspicious streaming figures. The company reserves the right to judge, at its own discretion, whether a release exists mainly to game the system. This too takes effect in mid-July.

The detail most likely to draw attention is money. From the day the policy was published, wholly AI-generated music will not earn royalties on Tidal. The company acknowledges the argument that some AI work, particularly material built on properly licensed models, might eventually deserve a share. It does not settle that debate. Instead it states a priority, which is that royalties should go to music written, performed and produced by people. Until the licensing picture becomes clearer, Tidal says it will not knowingly pay out on tracks it considers entirely AI-generated.

The same rules apply to Tidal Upload, the route independent artists use to place their own work on the service and earn from it directly. Because that material sits alongside the main catalogue, it faces the same checks. Anything Tidal believes to be wholly AI-generated will be tagged and shut out of direct-to-fan monetisation.

What Tidal has produced is less a ban than a set of boundaries. AI music is allowed in, labelled on the way, and kept away from the payment pool for now. The approach reflects a platform trying to hold a position while the ground keeps shifting, and while rightsholders and AI companies work out licensing arrangements that do not yet exist in any settled form.

For artists watching the streaming economy contract, the monetisation clause is the part that matters. It draws a clear line between human authorship and machine output at the point where revenue is decided. Whether that line holds as detection tools improve, and as licensed models complicate the question of what counts as AI-generated, is the test the policy now faces.

by Damion Pell

• Share
• Tweet
• Pin it
• Share
• Email
• More