25 September 2026 — The AI music industry has entered a new phase. The technology is improving rapidly, but this week's biggest story is about something much more fundamental: what happens when an AI music model carries knowledge from one generation of technology into the next?
The AI music industry is moving fast.
Very fast.
But this week has produced a development that could have consequences far beyond another improvement in music generation.
Universal Music Group and Sony Music Entertainment have launched a second major legal action against Suno, this time identifying 60,202 sound recordings in the new case.
The timing is significant.
Suno launched its new v6 generation only days earlier, with licensing partnerships involving Warner Music Group, BMG and Believe.
The result is an extraordinary collision between two versions of the AI music future.
One future says AI music can move forward through licensing, partnerships and controlled access to music catalogues.
The other asks whether a new model can genuinely escape the legacy of what previous models learned.
And that is where this week's story begins.
On 18 September, Universal Music Group and Sony Music Entertainment filed a second copyright lawsuit against Suno in the US District Court for the District of Massachusetts.
The new complaint identifies 60,202 sound recordings that the labels allege were copied without permission for AI training. The labels also argue that the alleged infringement extends beyond those recordings.
The sheer scale of the number is striking.
But the more important issue may be model lineage.
The labels are not simply arguing that Suno's older models were trained using allegedly unauthorised recordings.
They are arguing that the knowledge and behaviour developed by earlier models could have been carried into the company's new v6 models through processes including user feedback, previous model outputs and knowledge distillation.
In other words, the question is no longer simply:
“What data trained the new model?”
It is becoming:
“What did the new model inherit from the models that came before it?”
That is a much more complicated problem.
And it could become one of the defining legal questions of generative AI.
Suno disputes the labels' position.
In a statement released on 22 September, the company described the latest claims as fundamentally flawed and said its v6 models were developed using licensed material alongside interactions with its user community, including creations and preference signals.
That creates an important distinction between an allegation and an established fact.
The lawsuit contains the record labels' allegations.
Suno has presented a different account of how v6 was developed.
No final court ruling has determined those disputed questions.
For producers watching this unfold, however, the underlying issue is fascinating.
If a company builds a new AI model using properly licensed material but the new model also inherits information from an earlier model, does the new model have a clean legal starting point?
That question is now being tested in court.
Some coverage of the new lawsuit has focused on a potential damages figure of more than $9 billion.
That number needs to be treated carefully.
Music Business Worldwide notes that US copyright law permits statutory damages of up to $150,000 per work for willful infringement. Multiplying that maximum by 60,202 recordings produces a theoretical figure of just over $9 billion. That is not a damages award and does not mean Suno has been ordered to pay anything close to that amount.
The important story is therefore not the headline number.
It is the scale of the dispute.
More than 60,000 recordings are now specifically identified in the new case.
That demonstrates how far the legal fight has expanded since the original AI music lawsuits began.
Here is where the story becomes particularly interesting.
Suno has not simply launched v6 into the market without music-industry relationships.
The company has partnerships involving Warner Music Group, BMG and Believe.
At the same time, Universal Music Group and Sony Music are pursuing a second lawsuit.
So the music industry is not speaking with one single voice about AI.
Some rights holders are pursuing licensing relationships.
Others are continuing litigation.
And some companies are effectively doing both at different times and in different circumstances.
That tells us something important.
The music industry's relationship with AI is becoming increasingly commercial and strategic.
The argument is no longer simply:
AI versus the music industry.
It is becoming:
Which AI companies, which data, which licences, which artists, which business models and under what conditions?
That is a far more complicated — and potentially more sustainable — industry.
And this week brought another development that expands the AI copyright debate beyond recorded music.
On 24 September, sound designers, recordists, mixers and sound-effects library companies launched the Professional Sound Alliance.
The coalition says it wants fair licensing for professional sound design and sound effects used in film, television, games and AI training.
The organisation identifies two forms of unauthorised use: what it calls Data Piracy, involving the scraping of sound libraries into commercial AI models, and Sync Piracy, involving unlicensed use of professional sound libraries in productions.
This is significant because it demonstrates how the AI debate is spreading beyond songs and recordings.
Professional sound contains years of human work.
Field recording.
Editing.
Noise reduction.
Classification.
Metadata.
Curation.
Sound design.
Mixing.
And creative judgement.
AI training does not only involve famous songs.
It can involve the enormous ecosystem of professional audio that sits underneath film, television, games and music production.
The Professional Sound Alliance is therefore another sign that creators across the audio industry are beginning to ask the same fundamental question:
If your work helps train a commercial AI system, where is the permission and where is the compensation?
Meanwhile, Deezer is continuing to push aggressively into AI identification.
The company received the Billboard France Innovation Award this month for its AI-music detection technology.
Deezer says its system can identify fully AI-generated tracks with 99.8% accuracy. It also reports that nearly 90,000 fully AI-generated tracks were being delivered to Deezer every day in June 2026, representing more than half of new uploads at peak levels.
But the most interesting statistic is what happens after detection.
Deezer says AI-generated music accounts for only around 1–3% of plays on its platform despite representing a much larger share of uploads.
The company attributes this partly to measures including excluding fully AI-generated tracks from algorithmic recommendations and editorial playlists, labelling albums containing AI-generated music and demonetising streams it identifies as fraudulent. Deezer says fraudulent activity can account for up to 85% of plays associated with this AI-generated content.
That paints a fascinating picture.
AI may be capable of producing music at enormous scale.
But scale does not automatically translate into audience attention.
This could become one of the most important developments for independent AI producers.
Generating music is getting easier.
Getting people to listen is not.
If streaming services become increasingly effective at filtering out fraudulent, low-value or purely automated material, the advantage of producing thousands of tracks disappears.
The producer is pushed back toward the fundamentals:
Good music.
Strong identity.
Good artwork.
Interesting songwriting.
Consistent releases.
Audience development.
And genuine reasons for people to come back.
In other words, AI may actually make artist development more important, not less.
Spotify appears to recognise this.
On 17 September, the company launched Fresh Finds Forward, a programme designed to support approximately 10,000 emerging independent artists each year.
The programme offers eligible artists access to resources including studio time for US-based artists, early access to selected Spotify tools, opportunities around touring and music-video partnerships, networking opportunities and selected creative software benefits.
Spotify says Fresh Finds playlists have helped more than 80,000 artists across 127 countries and that, since 2024, those playlists have generated more than 128 million new artist discoveries.
The timing is interesting.
Spotify is simultaneously increasing transparency around AI artist identities while investing in emerging independent human artists.
Its AI Persona system is designed to identify artist profiles whose public identities appear to be AI-generated, while its Fresh Finds Forward programme is putting resources behind developing human artists.
That does not mean Spotify is rejecting AI music.
It means the platform is increasingly distinguishing between AI as a creative tool and AI as an artist identity.
That distinction could become central to the streaming ecosystem.
There is a useful way of looking at this.
A producer can use AI for:
Composition.
Arrangement.
Sound design.
Vocal experimentation.
Mixing.
Mastering.
Idea generation.
Performance assistance.
None of those things automatically determines who the artist is.
The artist is still the person making the creative decisions.
That is why the phrase “AI-assisted music” may become increasingly useful.
It describes a process rather than pretending the technology itself is necessarily the artist.
A human producer directing an AI system is one creative relationship.
An entirely synthetic artist identity is another.
An automated content operation generating thousands of tracks is something else again.
Treating all three as simply “AI music” misses the important differences.
For independent producers, the lesson from this week's news is straightforward.
Do not build your entire career around the ability to generate music quickly.
That capability is rapidly becoming universal.
Build around what you can decide.
Your musical taste.
Your songwriting.
Your arrangement choices.
Your sound.
Your visual identity.
Your artist name.
Your story.
Your audience.
And your ability to turn technology into something recognisably yours.
The technology is becoming cheaper.
Creative identity is not.
The professional AI producer also needs to become increasingly comfortable with documentation.
Know what tools you used.
Know what their terms allow.
Know whether commercial use is permitted.
Keep your project files.
Keep original lyrics and recordings.
Keep track of your creative contribution.
Maintain accurate metadata.
Understand your distributor's rules.
And do not artificially manipulate streams.
The streaming platforms are becoming more sophisticated.
The rights holders are becoming more organised.
AI detection is improving.
Legal scrutiny is increasing.
The independent producer needs to keep up.
The most interesting development of 2026 may eventually turn out to be the shift from music generation toward creative infrastructure.
AI companies are building models.
Record companies are licensing catalogues.
Streaming services are developing detection systems.
Distributors are facing increased scrutiny.
Sound libraries are demanding licensing protections.
Artists are building new identities.
And independent producers are trying to work out where they fit.
The producer who only thinks about generation is looking at one small part of the system.
The producer who thinks about creation, ownership, identity, distribution and audience is looking at the entire pipeline.
That is a much more powerful position.
AI has dramatically lowered the technical barrier to making music.
It has not eliminated the need for professionalism.
In fact, the opposite may be happening.
The easier it becomes to create something, the more important it becomes to demonstrate that what you created has a genuine identity and purpose.
Anyone can generate.
Not everyone can build an artist.
Anyone can upload.
Not everyone can build an audience.
Anyone can produce a track.
Not everyone can create a catalogue that people want to revisit.
That is where the independent producer still has enormous creative territory.
** COMING SOON **
PLEASE WATCH OUT FOR SOME GROUND BREAKING NEWS FROM SELFSOUND.COM.
STAY TUNED!
This week's developments also reinforce an idea that has been central to Selfsound.com: an independent producer's digital identity should not be treated as an afterthought.
Streaming platforms are becoming more sophisticated about AI identity.
Distributors are under greater scrutiny.
Rights holders are becoming more demanding about provenance and licensing.
And the AI music ecosystem is becoming increasingly dependent on databases, platforms and centralised services.
That makes Decentralized Production Domains (DPD) particularly interesting.
The concept explores a different approach to artist identity — one in which the producer can establish a persistent Web3 presence connected to their music and creative work.
It does not replace streaming platforms.
It does not mean abandoning distributors.
It means thinking beyond them.
The question becomes:
If the platforms change, where does your artist identity remain?
If your distributor changes its policies, where is your independent digital home?
If your music exists across multiple services, what connects it all together?
Those questions are becoming increasingly relevant in an AI music industry where the rules are changing almost every week.
Explore Selfsound's Decentralized Production Domains:
Selfsound.com — Decentralized Production Domains
And explore the Web3 Audio decentralised production domain:
The AI music story has changed again.
This week is not really about another AI model becoming better at writing songs.
It is about trust.
Universal and Sony are challenging Suno over 60,202 recordings and arguing that the company's newest models cannot simply escape the legacy of earlier training.
Suno disputes those claims and says v6 was developed using licensed material alongside interactions and signals from its community.
Deezer is demonstrating that AI detection and filtering can operate at enormous scale.
The Professional Sound Alliance is expanding the licensing debate into professional sound effects and sound design.
Spotify is simultaneously labelling AI-generated artist identities and investing in emerging independent artists.
And the entire industry is moving toward a future where provenance, identity and transparency are becoming part of the product.
The old question was:
“Can AI make music?”
That question has already been answered.
The new questions are much harder:
Who made it?
What was it trained on?
Was the material licensed?
Who is the artist?
Who owns the result?
Are the streams genuine?
And does the music actually matter to somebody?
For independent producers, this is not necessarily a reason to step back.
It is a reason to become more professional.
Use the technology.
Experiment.
Create.
Push boundaries.
But build something that has your identity running through it.
Because the future of AI music will probably contain more music than the world has ever seen.
The scarce resource will not be music.
It will be attention, trust and identity.
And that is where the serious independent artist has an opportunity.
THE FRIDAY NEWS BLAST — 25 September 2026
AI Music. Real Creativity. The Industry Is Changing.
