The charts just drew a line through synthetic music
Position on the ai music charts has always been the proof point a release campaign is built on. As of this week, that proof point comes with a provenance test attached.
A provenance test lands on ai music charts#
A fox does not argue with a fence. It finds where the fence ends. Anyone doing music marketing this quarter should be doing the same reading, because a new fence went up on 30 July and it runs straight through the release plan.
IFPI rolled out global principles for the eligibility of recordings developed using AI in official music charts worldwide. Read the scope carefully, because it is staged rather than universal: IFPI has applied the principles to the ai music charts it directly manages, covering Latin America, the Middle East and North Africa, Africa and South East Asia, and says it is working with its national group network to expand ai music charts eligibility across an additional 20 or more official chart programmes. Circle Chart, ARIA Charts and the Official Aotearoa Music Charts sit in that second group, so they are coming rather than live.
Three core criteria decide eligibility, and every music marketing plan now has to answer all of them. In IFPI's words, the AI service used must be "properly authorised and lawful"; the track must be "substantially human made"; and it must not "raise stream or chart manipulation concerns".
The disclosure principle is the marketing one#
Those three are legal, creative and integrity tests. Three further principles sit in IFPI's notes to editors, including compliance with applicable laws covering copyright, related rights and personality rights, and it is the last of the six that lands squarely on marketing: the use of generative AI must be "appropriately signalled to consumers on downstream services". That is a music marketing obligation,
Victoria Oakley, CEO of IFPI, framed the intent plainly: "Official music charts do more than just track sales; they celebrate human artistry and endeavour."
| Principle | Stated test | Who owns it in practice |
|---|---|---|
| Core 1: lawful tooling | The AI service used is properly authorised and lawful | Legal and business affairs |
| Core 2: human authorship | The track is substantially human made | A and R, producer, artist |
| Core 3: chart integrity | It does not raise stream or chart manipulation concerns | Label and digital distribution |
| Sixth principle: disclosure | Use of generative AI appropriately signalled to consumers on downstream services | Marketing and digital distribution |
The labels asked for this themselves#
This did not arrive as regulation imposed on an unwilling industry, which is why music marketing teams cannot wait it out. It arrived as an industry proposal, which makes it considerably harder to lobby away and much more likely to stick.
On 29 July 2026, eleven major and independent label groups proposed the ai music charts eligibility rules: Sony Music, Universal Music Group, Warner Music Group, Believe, BMG, Concord, Dirty Hit, Glassnote Records, HYBE Corp., Mom+Pop Music and Partisan Records. Their stated goal was to "establish a clear boundary between human-led creativity and works that are purely synthetic or generated by unauthorized AI models".
One of those core criteria deserves separate attention, because it is easy to read as somebody else's problem. Recordings must not "raise stream or chart manipulation concerns", and the making available of the recordings must not breach "the terms of the generative AI service used". Read that second one twice: the licence terms of the tool your producer used now decide eligibility for ai music charts.
Charts live now
Regions where IFPI has applied the principles to charts it directly manages: LatAm, MENA, Africa, South East Asia.
Further programmes
Additional official ai music charts programmes IFPI says it is working to expand the principles across.
Label groups
Majors and independents that proposed the framework on 29 July 2026.
Core criteria
Core eligibility tests, within a full set of six principles.
Independent confirmation of the proposal came from Billboard and Digital Music News on the same day, while Variety reported the appetite for scrutiny of AI tracks before they chart rather than after.
The commercial logic is not mysterious. Deezer reported in April 2026 that AI-generated tracks represented roughly 44% of new uploaded music. When nearly half of new supply is synthetic, a chart that cannot distinguish provenance stops being a signal at all, and a signal nobody trusts cannot sell a record.
Munich supplied the enforcement#
Rules without consequences are guidance. Two days after the label proposal, a German court supplied the consequence.
On 31 July 2026 the Munich Regional Court ruled against Suno in a case brought by GEMA. According to JUVE Patent, the case is 42 O 763/25 before the 42nd Civil Chamber, presiding judge Elke Schwager, and concerned six musical compositions. The court ordered Suno to cease the unauthorised reproduction of the protected works and to stop using them to train the AI model, and ordered disclosure of revenue connected to the infringements plus payment of damages.
Coverage from Music Ally and Variety confirms the outcome, and the Reed Smith analysis unpacks the dual finding on training and reproduction.
Now put the two events side by side, because together they redraw the music marketing brief. IFPI made lawful tooling an eligibility criterion for ai music charts on 30 July. A court found on 31 July that a major generative music service had been training on protected works without authorisation. A campaign that used such a tool is not merely awkward, it is arguably ineligible, and nobody in the marketing chain will have asked which model the producer used.
Why the industry can afford to be strict#
Rules of this kind get written when the incumbent business is strong enough to defend rather than desperate enough to dilute, and that shapes what music marketing can credibly promise. Universal Music Group reported on 30 July 2026, and the numbers explain the confidence.
UMG's half-year results record total revenue of €6,194 million, up 5.3% year over year and 10.8% in constant currency. Subscription revenue reached €2,670 million, up 9.0% and 14.5% in constant currency. Ad-supported streaming revenue was €728 million, up 2.4% and 8.3% in constant currency. Adjusted EBITDA was €1,310 million, down 1.9% but up 2.7% in constant currency.
One line in the release deserves its own paragraph for anyone building a music marketing strategy. Wholesale price increases contributed 3.5 percentage points to Q2 subscription revenue growth, partially offset by a 1.5 percentage point negative impact from market share headwinds. Quote both halves: it is a rare hard number on what subscription pricing actually contributes, and the offset is what stops it being a fairy tale.
The release also cites market share headwinds from Q1 2026 extending into Q2 2026, which is the honest other half. Coverage at Music Business Worldwide covers the quarter and the Downtown consolidation effect, while Music Ally published analyst scepticism on 3 August.
Deezer, meanwhile, reported on 28 July 2026 that its growing direct business "delivers sustained profitability and cash generation, enabling investment in future growth and AI leadership", per its H1 2026 results release. Read the guidance alongside it: the company confirms FY26 revenue at the same level as FY25, so this is investment funded from flat topline rather than from growth. A platform treating AI detection as a competitive investment rather than a compliance cost is still the clearest tell that provenance has become a product feature.
Five practical moves for music marketing teams#
The good news for music marketing is that none of this touches the creative. It touches the paperwork that follows the creative, which is where music industry marketing has always been weakest and where it is cheapest to improve.
Add one line to the stem delivery checklist: which generative tools were used, on which elements, under which licence. Asking at master stage is asking too late.
Store the actual terms of the generative service used, dated. The Suno ruling makes the lawfulness of the tool a live question, and a screenshot is worth more than a memory.
Make the AI-use flag a required field in your distribution metadata so it reaches downstream services. This is the third IFPI criterion, and it is an operations task.
Run one back-catalogue release through the three criteria as a dry run. You will find the gap in your own workflow before it costs you a place on the ai music charts.
If a record is substantially human made, that is now a claim with a governing body behind it. Put the studio, the players and the process in the campaign rather than saving them for the liner notes.
Provenance used to be a story you told after the record worked. It is now a condition of the record being counted at all.
The fifth move is the one with genuine upside. For a decade the industry has treated the making of a record as backstage material. A rule that rewards human authorship turns that backstage material into the campaign, and the artists who document their process now will look prescient rather than defensive in a year.
That repositioning is the work behind folkfox music industry marketing, and it draws on the storytelling discipline in content marketing.
What a label should count from here#
Eligibility is binary, which makes it unusually easy for a music marketing team to measure. You either can evidence the three criteria for a release or you cannot.
Start with disclosure metadata, because it is the gap that a distributor can close for the whole catalogue in one change and the one no amount of goodwill fixes at submission time.
The number to take to the board#
Report the share of forthcoming releases with all three eligibility criteria evidenced in the file. It is a percentage, it is countable this week, and it converts a vague anxiety about ai music charts into a number that either goes up or does not. A label that cannot state it is not protected by the fact that nobody has asked yet.
Set the baseline before changing the workflow, the way a fox tests a stream with one paw rather than trusting its look. Take the next twenty releases, score them against the three criteria, and let the failure rate rather than instinct decide how much process to add. Most catalogues need one metadata field and one question at delivery, not a governance programme.
The wider lesson travels well beyond music marketing. Every category where generative tooling has flooded supply is heading for the same destination: distribution conditioned on provenance. Music simply got there first, and the same logic is already visible in paid social and in how brand strategy now has to account for what a machine can and cannot be allowed to say on your behalf.
Frequently asked questions#
What are the new rules for ai music charts?
IFPI rolled out global principles on 30 July 2026. The three core criteria are that the generative AI service used is properly authorised and lawful, that the track is substantially human made, and that it does not raise stream or chart manipulation concerns. Three further principles follow, including that AI use is signalled to consumers on downstream services.
Does using any AI tool disqualify a track from ai music charts?
No. The test is not whether AI was used but whether the recording is substantially human made, whether the tool was lawfully licensed, and whether it raises chart manipulation concerns. Purely synthetic recordings and those made with unauthorised models are the target, and a sixth principle asks that AI use be disclosed downstream.
Which labels are behind the ai music charts eligibility rules?
Eleven groups proposed the framework on 29 July 2026: Sony Music, Universal Music Group, Warner Music Group, Believe, BMG, Concord, Dirty Hit, Glassnote Records, HYBE Corp., Mom+Pop Music and Partisan Records. IFPI then rolled it out across its global chart network.
What did the Munich court decide about Suno?
On 31 July 2026 the Munich Regional Court ordered Suno to cease unauthorised reproduction of six GEMA-protected compositions and to stop using them to train its model, and to disclose related revenue and pay damages. The damages amount is undetermined, the judgment is not yet enforceable, and Suno intends to appeal.
How much new music is AI-generated?
Deezer reported in April 2026 that AI-generated tracks represented roughly 44% of new uploaded music. That figure predates the chart rules and is widely understood as the pressure behind them, since a chart that cannot distinguish provenance stops functioning as a signal.
What should a release campaign change first?
Distribution metadata. The principle requiring AI use to be signalled to consumers on downstream services is an operations task, not a statement, and most workflows have no field to carry it. Fixing that field covers the whole catalogue at once.
Read more on this topic#
ChatGPT Paid Ads for Music and Events
Where release budgets and music marketing spend go when ai music charts are no longer the only proof point worth buying against.
See the music playbookThe platforms just made AI slop a distribution problem
The same provenance logic arriving in social feeds, where the campaign for the record actually runs.
Read the paid social pieceAI generated advertising just lost the right to stay quiet
The EU disclosure regime landing the same week, covering the campaign assets rather than the recording.
Read the disclosure pieceYour best content does not live on your website any more
How to measure artist content that lives entirely on platforms you do not own.
Read the measurement piece
Campaigns that survive the ai music charts eligibility test?
folkfox builds music marketing for ai music charts that holds up under scrutiny: provenance evidenced, disclosure wired into distribution, and the human craft moved from the liner notes into the campaign.