All articles

A Policy for Fully AI Music in TIDAL: No Paying for Full-AI Tracks!

It is not uncommon for big streaming services to change their terms and conditions or implement something new to attract more clients or keep the current customers satisfied. However, on June 29, 2026, TIDAL launched a new policy related to the AI-generated content and the way such content is paid. I can assure you that this change in the terms and conditions of a particular streaming service will have consequences for artists, AI musicians, and listeners as well, which makes it the most significant streaming service policy change since the 1,000 Stream Royale on Spotify.

The Short Version

This is TIDAL’s new AI policy: Starting July 15, 2026, all tracks identified by the service as 100 percent AI generated will receive a visible AI label in the app, will earn zero royalties, and will be excluded from the possibility to monetize the content via direct to fan TIDAL Upload. All listeners of TIDAL will receive an option to disable playback of AI labeled recordings by changing their settings and thus eliminate it from personalized recommendations. Any AI music that impersonates a real artist, deceives the listeners, and relates to fraudulent streaming gets blocked or removed from TIDAL automatically. Thus, AI music is not banned by TIDAL, it is demonetized and labeled. Let me discuss what does that mean for you in detail.

What TIDAL Actually Announced

As TIDAL announced on June 29, 2026, TIDAL is going to adopt a position regarding the AI-generated music that is published on its platform as a separate page of AI policy. The document is quite explicit and clear, and, in my opinion, unusual for a corporate text. TIDAL defines the AI generated music as “music that is wholly or substantially generated by generative artificial intelligence”, and commits itself to holding this content to “a higher standard of content integrity”.

There are four pillars of TIDAL’s new AI music policy. First, TIDAL will accept AI generated music. The company makes it quite clear that it believes that artists should have the freedom to create AI music, and listeners should have the freedom to choose what they want to listen to. Second, TIDAL will identify and label AI generated music, by introducing an icon on the tracks identified as 100 percent AI generated starting mid-July. Third, AI music that exploits someone’s music, name, likeness, deceives the listeners, and relates to fraudulent activities gets refused or removed. Finally, and this point attracted media attention all over the world, TIDAL decides that wholly AI generated recordings will not be eligible for royalty attribution at all.

Robert Andersen, the head of TIDAL at its parent company Block, explains the motive behind this decision on his X account by stating that the service receives “an overwhelming amount of AI-generated music” from third party distributors. According to the coverage in Variety, TIDAL’s primary goal is to ensure that royalties flow to the original works created, written, and performed by humans. Additionally, the company states that the problem of licensed training data is still discussed, and that it may reconsider the policy regarding the monetization of AI assisted music if industry standards change. In other words, the door is not closed forever, but it is locked for now.

The Dates and Mechanics That Matter

If you upload music, you should mark these dates on your calendar. The policy is published June 29, 2026, it starts to be enforced from July 15, 2026, and the icon that indicates the AI music starts to appear in the app for the listeners from mid-July, only on the tracks identified as fully AI generated.

The listener side of this policy is quite well-designed, and I am saying this as the person who complained much about poorly designed features in the past. According to the information provided by TIDAL in its own support pages, you can turn off AI labeled recordings in your settings. Once the toggle is off, AI labeled tracks will not play, and will be removed from your personalized recommendations as you receive new ones every day. Album recommendations, however, may take up to a few days to update. During this period, you may see some AI labeled music, but it will be grayed out and unplayable. If you see a track that you consider to be fully AI generated, but is not labeled, you can report it via the support form.

On the creator’s side, TIDAL Upload, the self-service distribution tool for independent artists, operates under exactly the same terms as label delivered catalog does. If TIDAL considers a particular upload to be wholly AI generated, it will receive a label and will lose the ability to earn royalties or any tips or payments from fans. There is no separate treatment for the independent artists here, and I think this is the right decision. Otherwise, the loophole will be abused very quickly.

One more mechanism, which is mostly missed in most of the publications on the topic: TIDAL pushes the responsibility for identifying AI generated music to its distributors. As it is reported by Headliner Magazine, TIDAL expects that the content distributors identify AI generated content before it arrives to TIDAL, and TIDAL plans to enforce this rule. If you upload music to TIDAL via DistroKid, CD Baby, TuneCore or any other distributor, expect that soon there will be no optional boxes for AI disclosure, it will become a mandatory field.

Why TIDAL Decided to Demonetize AI Generated Music

Nowadays, almost every major platform has adopted a certain policy towards the AI-generated music. Each platform builds its policy based on different fears. Spotify built its policy to fight the problem of spam and impersonation. Deezer focuses on fraud detection, and Apple Music prefers self-disclosure. Bandcamp decided to prohibit AI music altogether. TIDAL is the first platform to introduce monetization as the primary ground for its decision, and I think that this is a very wise choice.

Let me explain why. Royalties on any streaming platform are paid from a shared pool. Every stream that goes to a spammer who uses a prompt and bots to generate thousands of songs an hour dilutes the pool, and reduces the earnings of working musicians. For TIDAL it is not only a problem of housekeeping, but an existential brand issue as the company positions itself as an artist friendly platform with better per stream payments than its competitors.

According to TIDAL’s editorial leadership, this policy does not mean that the platform is against the technology. Tony Gervino, TIDAL’s executive vice president and editor in chief, told Inc. that AI has a true value in creating music. The example he provided is an independent artist that needs a horn section in his/her song, but cannot afford to hire a band. This distinction is important, as using AI as an instrument during the creation of music is okay and unlabeled. Typing a prompt and uploading the result is what gets labeled and demonetized.

The Numbers Behind the Panic

The TIDAL’s policy may look drastic, but let me provide you with the numbers that explain it. Deezer is the only major platform that publishes some numbers related to the AI music, and these numbers are quite alarming. In its April 2026 announcement, Deezer reported receiving almost 75,000 fully AI generated tracks a day, which is approximately 44 percent of the daily uploads and more than 2 million AI tracks monthly.

The growth rate is quite frightening. The number of AI tracks reported by Deezer itself, and excellently summarized in the coverage by TechCrunch, looks like this: 10,000 a day in January 2025, 20,000 by April 2025, 30,000 by September, 50,000 by November, 60,000 by January 2026, and 75,000 by April 2026. In July 2026, Deezer announced that the share had surpassed 50 percent of all daily uploads. More than half of the new music uploaded to a major streaming platform is now generated by machines, eighteen months ago the share was one in ten.

Two additional Deezer numbers complete the picture. Despite the huge amount of new AI music that Deezer receives every day, AI tracks account for only 1 to 3 percent of total streams on the platform. Of those streams, 85 percent are detected as fraudulent and demonetized. Read those numbers, and the conclusion becomes obvious: almost nobody is listening to that AI-generated music. Almost all uploads are done in order to collect the royalty money from the platform. The numbers of TIDAL are not disclosed, but the company describes the volume as overwhelming, and there is no reason to assume that the platform’s inbox looks significantly different from the one at Deezer.

How the Other Platforms Handle AI Music

In order to understand TIDAL’s decision, it is necessary to take a look at the industry in general, as the industry divides into four camps now.

Spotify pays AI music the same way it pays regular music. The September 2025 policy update by Spotify covers three points: impersonation policy regarding unauthorized AI voice clones, a music spam filter to deal with the massive uploads of artificial short tracks, and DDEX standard support for the standardized AI disclosures in the credits. DDEX is a standard that allows specifying the places in which the AI was used, whether in vocals, instruments, or post production, instead of putting the AI label or not AI binary. Also, Spotify reveals that it has deleted more than 75 million spammy tracks in the last twelve months, which gives us an idea of the scale of the problem. But the key point is that a fully AI generated track that follows the rules is eligible for royalties on Spotify.

Deezer demonetizes fraudulent streams, not AI itself. Deezer labels fully AI generated tracks, removes them from algorithmic and editorial recommendations, stops storing hi-res versions of the AI-generated tracks, and demonetizes the streams of AI-generated music if it detects that they are fraudulent. So, an honest AI track with real listeners may theoretically earn on Deezer.

Bandcamp banned AI music. In January 2026, Bandcamp, in its Keeping Bandcamp Human policy update, prohibits uploading of the music generated wholly or in substantial part by AI, as well as any impersonation of other artists or styles using AI. According to TechCrunch, the enforcement of the policy is done via user reports and reviewed by the human team, and Bandcamp is going to remove the music on suspicion alone. It is the harshest line in the industry, but it is also the harshest to enforce.

TIDAL occupies the fourth position on the board. The AI-generated music is allowed, labeled and not paid for. Regardless whether it is fraud or no fraud, honest or dishonest, a fully AI generated track on TIDAL earns zero. It is a commercial decision of the platform regarding the type of the content it wants to fund, and this is why the Forbes analysis of TIDAL’s policy calls it the step that no other streaming service had made yet.

What This Means if You Are an Independent Artist

If you create music the old-fashioned way or use AI in the process of music creation alongside with human work, this policy is almost entirely beneficial for you with one caveat that I am going to discuss below.

First, the good news. Every dollar that does not go to prompt farms stays in the pool of royalties. TIDAL is already one of the better-paying services per stream, and preventing the dilution of the pool helps you. Also, the AI label protects your discovery surface area, as Deezer’s data proves that AI tracks are barely streamed when the listeners know about them. TIDAL’s opt out toggle ensures that the audience that hates AI music will not have to compete with a tide of synthetic music in its recommendations.

The caveat is false positives. AI detection is pattern matching, and the music that uses heavy quantization, samples, and loops may be mistaken for AI output. If your track is mistakenly labeled, you are losing royalties on it until the label is removed, and TIDAL’s dispute path currently goes through the label or distributor rather than artist’s account. I strongly recommend you to save session files, stems, and DAW project history of every track that you released. It is your proof if you have to prove that the track was made by you.

Also, I suggest you to disclose AI usage honestly in the credits of the track through the distributor. The use of AI for mastering or stem separation is considered as AI assistance, and no platform currently penalizes you for that. Hiding the use of AI does not bring you any benefits and creates the risk that the detection flags something.

What This Means if You Make Fully AI Music

Honestly, I will not try to be diplomatic with you: the business model of uploading prompt generated tracks to streaming services and collecting the royalties is dying, and TIDAL is the platform that just accelerated it. Bandcamp will not accept your catalog at all. Deezer will bury it and demonetize most of its streams. TIDAL will host it, label it and pay you nothing for it. Spotify still pays, but its spam filter and the fact that the tracks created via AI persona profile are excluded from Verified by Spotify program show the direction in which it goes.

If you create something valuable with the help of AI music tools, the path is not the royalties from the streaming services. It is sync licensing for content creators, direct sales on the platforms that allow it, commissioned work, and building the audience that knows exactly what you are doing. The creators that I have seen succeeding with the help of AI assisted music are quite open about their process and treat AI as just one of the instruments with their vocals, playing, and arrangement work.

I would also warn you about the workarounds. Now there are platforms that offer you a service of making AI-generated tracks undetectable to the classifiers of platforms. Trying to use them in order to insert demonetized content to the royalty pool is a deceptive and fraudulent activity that TIDAL’s policy explicitly mentions as a reason to refuse and delete your content. Platforms share fraud signals with distributors, so getting your distributor account terminated brings down all your catalog, including all legitimate releases. Not worth it.

The Loophole Everyone Is Pointing At

The sharpest criticism of the TIDAL’s policy is directed to the word wholly. The rules apply to the content that is 100 percent AI-generated, and as iMusician reports, the users have immediately found the loophole: what about a track that is 95 percent AI generated? One commenter stated that quite brilliantly by wondering whether layering yourself playing a triangle over an AI song would make the latter monetizable and unlabeled. Strictly speaking, according to the current policy, yes, it might.

TIDAL knows about the loophole. In the policy, it explicitly states that labeling will extend to substantially AI-generated content as the detection methods become more reliable, and calls the whole document “living” that will evolve with the technology. In addition, TIDAL’s support pages openly admit that considering the focus on wholly AI generated content means that some of the partially AI generated content remains unlabeled for now.

My opinion is that starting from the narrow focus is correct and wise. A platform that launches with demonetization of substantially AI-generated content has to give a strict definition of the term “substantially” on the first day of launch, and it will certainly make mistakes that will damage real artists. Focusing on 100 percent AI generated tracks means that the only people affected are those who contributed absolutely nothing human to the track. The triangle loophole is real, but its exploitation at scale requires actual human work on each track, and it destroys the economics of the content farming. The fraud business works as long as the generation is instant and free. Adding even ten minutes of human work to each track means breaking the economic model of spam completely.

The Detection Problem Nobody Has Solved

And here is the unpleasant truth under any AI music policy: it relies on detection technology that is imperfect. A Forbes analysis from July 6 pointed out that with TIDAL’s policy AI detection technology becomes critical infrastructure for the industry, while highlighting its main flaw. Detection is accurate in recognizing AI generated audio, but cannot recognize whether an artist legitimately used it or it is a fake artist using the tool to generate streams. The waveform does not contain any intent information.

In January 2025, Deezer created their patent pending detection solution and shared it with other platforms. They even provided a scanning tool for Spotify and Apple Music playlists. It is still unclear which detection solution TIDAL is using. Moreover, models keep developing in the direction of becoming more human-like, and according to a survey commissioned by Deezer, 97 percent of listeners could not differentiate AI music from a human-generated song. This is an arms race, and generators iterate faster than detectors right now.

For all the participants of the ecosystem, this means that there will be false positives and false negatives. Some AI tracks will get through without labels, hence the need for a reporting function. On the other hand, some tracks performed by humans will get labels. Neither of these issues invalidates the policy, but if someone tells you that detection is solved problem – you are getting sold something.

What Listeners Actually Get

I would like to give more attention to the listener experience in this policy, because this is the aspect I like the most, but it is usually skipped in articles. As TIDAL explains in the statement, one of the motivations for the policy came from subscribers who expressed concerns regarding listening to wholly AI generated music. The response was neither a ban nor a shrug, but informed choice.

You see a label, so you know what you are listening to. You get a toggle, so you can exclude it from your listening experience. Recommendations are adjusted accordingly. In case you really like to listen to AI generated ambient soundscapes when studying – no one prevents you. It is a much more honest agreement than on Spotify, where AI music is unlabeled in editorial adjacent playlists unless a distributor added a disclaimer, and more flexible than Bandcamp, where the decision has been made for the listeners already. For fifteen years streaming platforms were making decisions on behalf of their users via opaque algorithms. A visible tag and an off switch are just a small thing that stands for a completely different approach.

What I think will happen next

Here are my three predictions with various levels of confidence.

First, disclosure of AI use on the distributor level will become mandatory for the entire industry within a year. TIDAL is already trying to push identification up, Spotify has DDEX commitments from more than a dozen of labels and distributors, and once two big DSPs ask for the same data – everyone provides it automatically. High confidence.

Second, the line between wholly and substantially AI generated track will move. TIDAL already said that they will expand labeling in the process of improving detection, and pressure from artists who saw the triangle loophole discussed publicly will keep it on the roadmap. Expect a substantially AI generated track classification with its own rules by 2027. Medium confidence.

Third, licensed AI music will get its monetization opportunity again. TIDAL left space for reevaluation of AI assisted music generated on licensed models, and Udio is already signing licensing agreements with major labels and Merlin’s independent network is being built. Once the rights infrastructure is mature enough – I expect a classification of disclosed, licensed, partially AI music earning royalties through labels. Lower confidence on the timing, but the policy text was clearly prepared for it.

What I would do this week

Artist: log into your distributor, check whether your tracks have AI disclosure metadata, fill it in accurately, and start storing stems and project files for all your future tracks. Then scan your catalog on TIDAL and ensure that none of your tracks is labeled as AI by mistake.

Listener on TIDAL: go to Settings, look for the AI labeled recordings toggle and set it the way you actually want. Report obviously unlabeled AI tracks when you see them, because the reporting loop is a part of the improvement of the detection process.

AI music creator: stop building revenue plans on streaming royalties, be transparent about your process and put real human creative effort into anything you want to monetize somewhere.

Sources

Common questions

Does TIDAL ban AI generated music?

No. TIDAL's policy, published June 29, 2026, explicitly says artists should be free to create AI music and listeners free to choose whether to hear it. What TIDAL does is label tracks identified as fully AI generated, exclude them from royalty payments, and remove anything that impersonates real artists, deceives listeners or involves streaming fraud.

Do AI generated songs earn royalties on TIDAL?

Not if they are wholly AI generated. From July 15, 2026 those tracks are ineligible for any royalty attribution, and the same rule applies to independent artists using TIDAL Upload, who also lose tips and fan payments on labelled tracks. The company says it may revisit monetization of AI assisted music if industry standards around licensed training data change, but for now the door is locked.

How do I turn off AI music on TIDAL?

Go to Settings and switch off the AI labelled recordings toggle. Once it is off, AI labelled tracks will not play and will drop out of your personalized recommendations as they refresh, though album recommendations can take a few days to update and any lingering AI tracks appear greyed out. If you spot a fully AI track that has no label, you can report it through TIDAL's support form.

Does using AI for mastering or stem separation get my track labelled on TIDAL?

No. TIDAL's editor in chief gave the example of an independent artist using AI for a horn section they could not afford to hire, and described that kind of use as an instrument in human creation, which stays unlabelled and monetized. The label targets tracks where someone typed a prompt and uploaded the result with no human contribution.

What is the TIDAL AI policy loophole?

The rules apply only to content that is 100 percent AI generated, so critics immediately asked whether adding a single human element, like a triangle part, over an AI track would keep it monetized. Strictly speaking it might. TIDAL acknowledges this, calls the policy a living document, and says labelling will extend to substantially AI generated content as detection improves. The article argues the narrow start is wise because even ten minutes of human work per track breaks the economics of content farming.

What if my human made track is wrongly labelled as AI on TIDAL?

This is the main risk for real artists, since detection is pattern matching and music built on heavy quantization, samples and loops can trigger it. A false label means lost royalties until it is removed, and the dispute route currently runs through your label or distributor rather than your own account. Keep session files, stems and DAW project history for every release as proof, and disclose any genuine AI assistance honestly in your credits.

Why did TIDAL decide to stop paying for AI music?

Because royalties come from a shared pool, and every stream that flows to prompt generated spam dilutes what human musicians earn. TIDAL's head at parent company Block said the service receives an overwhelming volume of AI music from distributors, and the company positions itself as an artist friendly platform with better per stream payments, so protecting the pool is a brand issue as much as housekeeping.

How much AI music is being uploaded to streaming platforms?

Deezer, the only major platform publishing figures, reported nearly 75,000 fully AI generated tracks a day in April 2026, around 44 percent of uploads, rising past 50 percent by July 2026 from just 10,000 a day in January 2025. Yet AI tracks account for only 1 to 3 percent of Deezer's streams, and 85 percent of those are detected as fraudulent, which suggests almost nobody is listening and most uploads exist to harvest royalties.

How do Spotify, Deezer and Bandcamp handle AI music compared to TIDAL?

Spotify still pays fully AI tracks that follow its rules, while filtering spam, removing unauthorized voice clones and supporting DDEX disclosure in credits. Deezer labels AI tracks, removes them from recommendations and demonetizes only streams it detects as fraudulent. Bandcamp banned wholly or substantially AI generated music outright in January 2026. TIDAL is the first to make monetization itself the dividing line, paying nothing for fully AI tracks regardless of whether fraud is involved.

Can I still make money from AI generated music?

Not reliably through streaming royalties, which the article says is a dying business model now that Bandcamp refuses it, Deezer buries it and TIDAL pays nothing. The paths that remain are sync licensing for content creators, direct sales where allowed, commissioned work and building an audience that knows your process. Tools that make AI tracks undetectable count as fraud under TIDAL's policy and can get your whole distributor account terminated.