Tidal to Tag Fully AI-Generated Music and Block It from Royalty Eligibility

Tidal has introduced a policy that automatically tags wholly AI-generated music uploaded to its platform and bars it from earning royalties, according to Music Business Worldwide (published 29 June 2026).

The move draws a hard line between human-made music and output generated entirely by AI systems — a distinction that streaming platforms have so far handled inconsistently, or not at all. Under the policy, content flagged as fully AI-generated will carry a visible tag and will be excluded from royalty pools. Tidal separately confirmed that music uploaded to the platform will not be used to train AI models, a commitment Digital Music News reported in March 2026.

The practical scope here is worth spelling out. Royalty pools on streaming platforms are shared: every stream of synthetic filler content dilutes the per-stream rate available to human artists. The volume of AI-generated audio — loops, background music, ambient tracks uploaded at scale — has been a documented source of friction between platforms and rights holders for several years. Tidal's policy directly attacks the economic incentive to flood catalogs with machine-generated content.

Tagging is the other half of the equation. A machine-readable label on AI-generated tracks enables downstream filtering by distributors, PROs, and potentially DSP recommendation systems. Whether those downstream actors will act on the tag is an open question, but the infrastructure for enforcement at least exists once labeling is in place.

The training-data commitment operates on a separate axis. Tidal's catalog — curated toward higher-fidelity, often lossless audio — would be a meaningful dataset for any organization building or refining audio generation models. A blanket prohibition on using uploaded content for model training, if enforceable, forecloses one route by which platform catalogs have been harvested without rights-holder consent.

Taken together, the two commitments address what have been the two most immediate complaints from professional musicians about the current AI-audio landscape: synthetic tracks eroding royalty income, and their recordings being ingested to train the very systems competing with them.

The harder question is enforcement. Detecting "wholly AI-generated" audio is a non-trivial classification problem. Watermarking standards for AI-generated audio — such as those being developed under the C2PA (Coalition for Content Provenance and Authenticity) framework — are still maturing, and self-declaration by uploaders is an obvious weak point. Tidal has not, based on available reporting, disclosed which detection methodology or provenance-verification toolchain it is applying. That gap matters. A policy that relies primarily on uploader honesty will produce different outcomes than one backed by spectral fingerprinting or embedded provenance signals.

The training-data prohibition raises its own compliance questions. Restricting third-party use of platform content requires auditable controls and, in adversarial cases, the willingness to litigate. Contractual prohibitions are a starting point, not a ceiling.

Looking at where this sits in the broader platform landscape: Tidal is not the largest streaming service by subscriber count, but it has historically positioned itself as the artist-aligned option — the platform where royalty rates and audio quality were points of differentiation. Extending that positioning into AI policy is consistent with the brand logic. Whether larger platforms with substantially bigger catalogs and more complex licensing structures follow is a separate question, and the commercial pressures they face are not identical.

What Tidal has done is set a documented, public baseline — a policy that can be measured against actual catalog data over time. For artists and distributors negotiating with platforms, having a named reference point for what AI-content policies can look like is itself useful, regardless of how other services respond. The music industry has been asking platforms for concrete commitments on AI. One platform has now provided two of them in writing.