Tidal's AI-label rule draws a line, but the royalty question is still open

Tidal told the music industry on 29 June 2026 that any track uploaded to its catalogue must be labelled as human-made, AI-assisted or fully AI-generated, and the platform said it will not distribute music to streaming services that fail to adopt a comparable standard. The Norwegian-American streamer, owned by a Square-blessed successor entity since Jack Dorsey's 2021 Block acquisition closed, framed the move as a transparency play for listeners. For an artist economy already arguing about attribution, sampling and credit, the announcement landed on a fault line that has been cracking for at least a year.

The label rule is, on its face, modest. Three buckets, a tagging obligation, a downstream-compliance trigger. What it does not do is solve the problem it is being marketed to solve: who gets paid, and on what basis, when a model trained on a back catalogue is asked to write something that sounds like the people it learned from. Tidal has drawn a line, but the line is around disclosure, not royalties. The harder question sits one headline down.

The announcement, as far as it goes

According to Variety's 29 June 2026 write-up of the policy, Tidal will require metadata on every new upload specifying whether the recording was made by a human artist, made with AI assistance, or generated entirely by AI. Tidal said the tags are intended to surface provenance to listeners and to give rights-holders a clean ledger of what is and is not synthetic. The platform also said it will refuse to distribute catalogue to partner services that do not match a comparable labelling threshold, a clause that, if enforced uniformly, would put pressure on every aggregator sitting between bedroom producers and the major streaming storefronts.

That is the policy as published. It is a disclosure regime, not a licensing regime. A track tagged "AI-assisted" still streams at the same royalty rate, under the same pro-rata pool, as a track tagged "human-made." Nothing in Tidal's statement changes how the money is divided. What changes is the paper trail.

Why artists are still angry

The disclosure framework does not address the central grievance now running through the artist community: that generative models have been trained on copyrighted recordings without consent, that the outputs of those models are competing with the originals in recommendation feeds, and that the royalty cheque on a fully human record and the royalty cheque on a model-generated facsimile will, under the current pool, look the same. The Human Artistry Campaign, the SAG-AFTRA-adjacent artist coalition, has been pushing for enforceable consent rules and for a separate royalty track for synthetic material; Tidal's buckets are downstream of that fight, not a replacement for it.

Major labels, for their part, have negotiated bespoke licences with a handful of model providers, and the terms of those licences are confidential. Indies and mid-tier artists have no equivalent leverage. A tagging regime, in that environment, functions as a sorting mechanism. It tells the market which streams are which. It does not tell the market which streams are paid at which rate.

The technical gap

There is also a question about how the rule will be enforced. Tidal has not, in the public version of the announcement, specified the detection mechanism. AI-generated vocals can be re-recorded, layered and mixed until a classifier returns a low confidence score. AI-assisted sessions can be made to look human in the metadata if the producer chooses. A tag is only as reliable as the upload portal that requires it, and the history of self-attested compliance in music is not encouraging: distributor terms-of-service have been signing off on splits for years, and disputes still end up in court.

What that means in practice is that Tidal's rule is a floor, not a ceiling. It is the minimum standard the platform will accept, and it leans on the uploaders' honesty. Major-label catalogue, with its legal departments and audit rights, is likely to comply cleanly. The long tail is where the labels drift.

Where the royalties actually go

The pro-rata streaming pool divides subscriber revenue across streams in proportion to total plays. A model-generated track that racks up plays under an unverified tag draws from the same pool as a human recording, and the per-stream rate is the function of total streams in the period, not the composition of those streams. The case for a separate synthetic-content pool, or a reduced rate for AI-generated streams, has been made most loudly in the US by the Recording Academy's policy arm and, in the UK, by the Musicians' Union. Neither has yet produced a model the major DSPs have agreed to implement.

Tidal's announcement does not move that needle. It is reasonable to read the move as positioning: Tidal wants to be the platform that took the disclosure step first, and it wants the credit for refusing to feed catalogue to competitors who refuse to follow. The royalty architecture is a different fight, and the labels know it.

What to watch next

Two dates are worth circling. The first is the next round of CRB-era Phonographic Performance Limited negotiations, where the rates paid for public performance in the UK get reset on a multi-year cycle and where any separate synthetic-content rate would have to be argued. The second is the first major DSP response to Tidal's distribution clause: whether Spotify, Apple Music, Amazon Music and YouTube Music publish matching metadata rules, or whether the clause functions as a commercial wedge. A third wildcard is the model providers themselves, several of whom have begun publishing "licensed training corpus" disclosures, in part because platforms are starting to ask.

The 29 June announcement is real, and it is consequential at the disclosure level. It is also, in the part of the argument that actually decides who gets paid, a press release. The royalty question stays open, and the next move is not Tidal's.

Sources: Variety, 29 June 2026; Wikipedia, Tidal (service); Recording Academy policy commentary; Musicians' Union statements on synthetic content, 2025–26; Human Artistry Campaign, 2025.