“My baby, My Dabby” was written 2 decades ago at the age of 19.

It’s just a silly song that actually was freestyle’d one summer night during a smoke session then later re recorded with WitchDoctor on Bass and doing backup vocals. So the original recording’s guitar, lead vocals, bass, and backup vocals are still driving the mix. Witchdoctors odd yet strangely fitting vocals during the instrumental section.

It’s just a silly song, I imagined myself as a Lousiana Boy whose woman ran off on him to New Orleans.

I only used Bandlab in this one and I added piano, horns, violin, guitar and more backup vocals.

The only AI used was in Bandlab and it was Audio to midi on my OG guitar to create more of a rhythm section, I “voice changed” one of my vocal takes to “Midnight Poet” but kept it subtle in the mix that you can’t really hear it, and the percussion I actually played on my fat beer belly and then Audio to midi’d it into drums lol. I did leave one section of the actual belly slapping though.

NO SUNO was used in this at all. Frankly I don’t consider the AI that was used in Bandlab to be “Generative” as it’s merely a function of Midi.

Visually you are seeing on the TV, An avatar of myself and the Witchdoctor ages 18-19 that I based on real images of us then I animated them with motion sync by acting out the parts myself. Although I put us in a more fitting hillbilly wardrobe. Witchdoctor would have never worn overalls but I am ashamed to say I did wear muscle shirts back then (cringe)

This is the second version of the song I posted, the first version was a mix between suno and the Original and this one has NO SUNO.

I spent WAY too much time on this for such a silly song.

Penny for ye thoughts?

I think the coming changes around fully AI gen music on DSPs are going to be much more strict than simply adding an “AI” label.

My guess is that, over time, DSPs and distributors will increasingly separate 100% AI generated music from human created music, with various shades of grey in between accounting for AI Assisted works.

In a nutshell here is what I believe is and should be coming for Fully Gen AI on DSP’s

-Upload Caps on the number and frequency an individual account can distribute

- No or heavily limited Algo boosting or pickups

- No placements on mainstream editorial playlists

- Dedicated AI playlists or discovery spaces

- Different monetization rules, higher thresholds and lower payouts, or no monetization at all

-But the biggest change and most important one for the future of music is NO CONTENT ID protection 

Fully AI generated music generally isn't copyrightable. And content Id exists to protect copyrightable works. It makes no sense that it receives Content Id.

Yes, there can obviously be copyrightable elements, if a human contributes original lyrics or other meaningful creative material, but trying to determine and enforce those individual components at scale is unrealistic for Streamers or DSP’s to manage.

Perhaps more can be done on Suno's end to provide receipts of user inputs?

Perhaps something can be done to address this nuance, but how will they know who is telling the truth about the lyrics vs who is lying.

Content ID exists to protect works in which someone actually has enforceable rights.

 Simply generating a piece of music, uploading it, and getting a fingerprint into a rights management system shouldn't give someone the ability to claim exclusive control over that material virtually indefinitely.

And this problem is only going to get worse as AI music scales.

Imagine 10 or 20 years from now, when AI farms and users have generated billions of songs. There is going to be an enormous amount of overlap sonically, even lyrically.  

Imagine some kid in 20 years sits down with a guitar, writes an original song, records a simple demo, and uploads it to share with the world, only to have it automatically claimed because an AI music farmer generated something too similar 20 years earlier and dumped it into a rights management system without giving it a second thought, or hell even have listened to it himself before dumping it on DSPs

That kid should not have to compete with billions of non copyrightable machine generated works for the right to distribute his own music. 

WE ARE ALREADY AS A SUNO COMMUNITY EXPERIENCING THIS. HOW MANY POSTS DO YOU SEE ON HERE OF PEOPLE'S OWN ORIGINAL SONGS BEING BLOCKED FROM SUNO?

That's why I think fully AI gen music should ultimately be treated differently when it comes to automated copyright enforcement.

It can exist. It can be distributed. People can listen to it, and there can be appropriate ways for people to monetize it.

But no Content Id protection for fully generated songs

Otherwise, AI users risk effectively robbing the creative commons through sheer volume.

In that sense, I think it's somewhat akin to people sitting on URLs in the early days of the internet.

They could register enormous numbers of domains simply because they had the ability to do so, effectively reserving pieces of the emerging internet without necessarily creating anything with them.

Mass AI music generation could create a similar form of creative land grabbing: generate millions of works at virtually no marginal creative cost, get them distributed and fingerprinted, and potentially occupy enormous amounts of musical territory that future human creators might independently arrive at.

The irony is that AI is supposed to “democratize” music creation (so it’s said)

But if we allow mass generated AI catalogs to accumulate automated enforcement power, we could end up creating a system where the machine generated past has priority over the human created future.

This is my second installment of my Pushback against the AI Music "community" and it will focus on why NO ONE, not just the grifters and spammers, should be uploading their fully generated songs to DSPS.

For anyone new to this sub I am TOP MOD, this is push back from your "own side".

(for you Aunties, yes I used AI to organize my thoughts, these are points in your favor so be grateful you are reading this in a Pro AI sub. If you demand I write out everything by hand I bill $75 an hour for post production work, let me know where to send my invoice. I don't owe you fucking shit, move on if you don't like the formatting)

To start: I DO NOT AND HAVE NOT Uploaded any of my songs through Commercial DSPS.

TLDR: The post argues that generative AI music skipped the essential cultural and institutional maturation stages—such as community building, artistic mastery, shared conventions, and legal frameworks—by jumping straight from creation to commercial distribution on streaming platforms. This rapid commercialization has led to platform flooding, "noise-to-signal" issues, public backlash, and diluted recognition for creators who put significant effort into their work. To prevent strict industry-wide crackdowns, the author proposes a voluntary "walk before you run" period where creators focus on experimentation, culture-building, and sharing tracks on non-commercial platforms rather than monetizing them on major DSPs.

If creators don't take curation and moderation seriously now, major streaming platforms and record labels will inevitably do it for them through draconian crackdowns.

The “walk before you run” metaphor

• The phrase captures the central problem:
  • AI music effectively tried to run before it learned to walk.
• The technology moved extremely quickly from:
  • Creation
  • → polished output
  • → commercial distribution
  • → monetization
• What was largely skipped was the intermediate period in which a new technology normally develops:
  • A community
  • A culture
  • Artistic experimentation
  • Shared conventions
  • New aesthetics
  • Standards of quality
  • Legal norms
  • Institutional infrastructure
  • A sense of identity and legitimacy
• The argument is therefore not simply:
  • “AI music shouldn't be commercial.”
• It is more specifically:
  • The technology entered commercial distribution before the surrounding ecosystem had time to mature.

The historical progression of music technology

The discussion proposes a general progression that major technological changes in music historically followed:

Stage 1 — The technology arrives

Examples include:

• Synthesizers
• Samplers
• Digital Audio Workstations (DAWs)
• MP3s

The technology initially provides a new capability.

Stage 2 — Cultivation and mastery

• Artists spend years experimenting with the technology.
• They discover what the tool can do.
• They push it beyond its obvious applications.
• Artists develop techniques and workflows.
• Distinctive artistic styles emerge.
• Communities form around the technology.
• Eventually, the technology becomes associated with particular sounds, scenes, movements, and cultures.

Stage 3 — Legal and institutional norms develop

• Society and the industry encounter problems created by the new technology.
• Rules and conventions develop in response.
• Examples include:
  • Sample-clearance practices.
  • Copyright frameworks.
  • Digital-rights systems.
  • Licensing structures.
  • Platform standards.
  • Industry norms.

Stage 4 — Commercial integration

• The technology becomes broadly integrated into the mainstream music industry.
• Commercial distribution catches up with the artistic and institutional ecosystem.
• The technology becomes normalized.

The argument that generative AI skipped stages 2 and 3

• Generative AI music developed extremely rapidly.
• Tools such as Suno and Udio allowed users to generate polished-sounding complete songs almost immediately.
• Instead of spending years developing an artistic movement around the technology first, users could immediately:
  • Generate songs.
  • Export them.
  • Upload them to streaming services.
  • Attempt to monetize them.
• In other words, the technology effectively went:

Technology → Commercial distribution → Monetization

instead of:

Technology → Cultural experimentation → Community → Artistic identity → Legal/institutional norms → Commercial integration

• This is the heart of the “walk before you run” argument.

Community and culture normally develop around new technology

• New music technologies historically did more than make existing music easier to produce.
• Artists used them to create new artistic languages.
• Communities formed around those new languages.
• Those communities developed:
  • Shared techniques.
  • Shared aesthetics.
  • Terminology.
  • Values.
  • Standards.
  • Influences.
  • Rivalries and debates.
  • Scenes.
  • Artists who became recognized as innovators.
• Over time, the technology became connected to an identifiable culture.

Why this matters for cultural legitimacy

• New artistic movements generally gain legitimacy through more than commercial success.
• They develop legitimacy through:
  • Artists experimenting.
  • Communities forming.
  • People developing mastery.
  • Audiences developing taste.
  • Critics discussing the work.
  • Artists influencing one another.
  • New techniques being invented.
  • A recognizable culture emerging.
• Commercial success eventually becomes one component of that ecosystem.

The DSP problem

• Traditional Digital Streaming Platforms (DSPs), such as Spotify and Apple Music, were built around existing models of music distribution.
• Their systems were not designed around a technology capable of producing essentially unlimited quantities of polished music with zero human friction.
• Generative AI dramatically reduces the cost and effort required to produce large quantities of tracks.
• A single person can generate and upload huge numbers of songs.

The resulting problem

• If enormous quantities of AI music enter DSPs:
  • The platforms become flooded with content.
  • Discovery becomes more difficult.
  • Listeners face more noise.
  • Existing artists have more competition for attention.
  • Royalty systems face additional pressure.
• This creates the noise-to-signal problem:
  • More content does not mean more meaningful music.
  • It can make meaningful music harder to discover.

The “heart vs. prompt” problem

• There can be a huge difference between two AI creators:
  • Someone who spends hours or days iterating, refining, directing, editing, and developing a song.
  • Someone who enters a three-word prompt and immediately uploads the result.
• Those processes can involve dramatically different levels of human involvement.
• Yet the final listener has no way to know which process produced a particular song.
• An automated platform may have even less ability to make that distinction.

Therefore:

• Human effort is largely invisible in the final track.
• A creator may genuinely feel that they created something meaningful.
• The listener simply sees:
  • “Another AI-generated song.”

The economic problem

• Traditional DSP royalty systems were built around a finite ecosystem of music producers and artists.
• Generative AI changes the scale of production.
• Traditional music production has natural limitations:
  • Human time.
  • Human labor.
  • Recording costs.
  • Studio time.
  • Creative energy.
• Generative AI eliminates those constraints.
• This makes music production potentially infinitely scalable—or at least dramatically more scalable than traditional human production.

This is one reason the technology creates problems that earlier musical technologies didn't create in exactly the same way.

The legal and institutional problem

• Because commercialization happened so quickly, legal and institutional systems are now reacting after the fact.
• The industry is dealing with questions surrounding:
  • Copyright.
  • Training data.
  • Ownership.
  • Rights.
  • Royalties.
  • AI disclosure.
  • Distributor policies.
  • Platform standards.
• Lawsuits and policy changes are occurring while the technology is already commercially active.
• In this framing, the infrastructure is being constructed after the technology has already entered the marketplace.

The public-friction problem

• Listeners and traditional creators have already become overwhelmed by enormous quantities of AI-generated content.
• Hell even inside the AI community people are exhausted with the sheer quantities of songs shared.
• This contributes to the perception of “AI slop.”
• The problem is partly that the normal cultural mechanisms for determining quality haven't had time to develop.

Those mechanisms historically include:

• Curation.
• Community validation.
• Artistic reputation.
• Mastery.
• Scene participation.
• Peer recognition.
• Audience taste.
• Cultural criticism.
• Generative AI effectively arrived at commercial scale before those mechanisms could develop around it.

Community is a form of cultural “gatekeeping”

• The argument isn't necessarily that gatekeeping is inherently good.
• Rather, communities naturally create mechanisms for determining:
  • What is interesting.
  • What is innovative.
  • What demonstrates skill.
  • What contributes something new.
  • What is derivative.
  • What is low effort.
• Those mechanisms help a new artistic medium develop standards.
• A mature community can distinguish:
  • Someone experimenting seriously with a new technology.
  • Someone producing enormous amounts of disposable content.
• AI music hasn't necessarily had sufficient time to develop these community-level distinctions.

The reputational danger

• Because low-effort AI music can be produced at scale, it dominates people's first impression of AI music.
• Listeners may therefore associate the entire technology with:
  • Spam.
  • Low effort.
  • Disposable content.
  • Commercial exploitation.
  • Lack of authenticity.
• This can unfairly affect creators who are actually putting significant effort into AI-assisted work.

The collective-action problem

• Individual AI creators may reasonably think:
  • “My song isn't slop.”
  • “I put enormous effort into this.”
  • “I deserve to distribute my music.”
• But if thousands or millions of creators make the same decision:
  • The aggregate effect can be very different from the individual case.
• This creates a classic collective-action problem:
  • An individual action may seem harmless while the cumulative effect creates systemic problems.

Systemic backlash (we are already seeing this)

• The discussion identifies several possible consequences of uncontrolled commercialization:
  • Platform restrictions.
  • Distributor bans.
  • Public backlash.
  • Stricter disclosure requirements.
  • Regulatory intervention.
  • Broader restrictions on AI-generated music.
• The irony is that responsible AI creators will be affected by restrictions created in response to irresponsible or excessive use of the technology.

The proposed middle ground

Rather than arguing that AI music should never exist commercially, the discussion suggests that the ecosystem needs time and infrastructure to mature.

Potential solutions include:

• Dedicated AI-music categories.
• AI disclosure requirements.
• Better provenance systems.
• Metadata indicating AI involvement.
• Distinct royalty structures.
• Platform rules specifically designed around generative music.
• Community standards.
• Better mechanisms for distinguishing high-effort AI-assisted work from mass-generated material.

Self-regulation as an interim measure

• Until those systems exist, I am proposing self-regulation.
• That could mean:
  • Continue making AI music.
  • Experiment with the technology.
  • Build communities around it.
  • share on non commercial sites (youtube, soundcloud, suno, reddit)
  • Develop new aesthetics.
  • Establish standards.
  • But hold back from flooding mainstream DSPs with commercially monetized AI tracks.
• The purpose wouldn't necessarily be to declare AI music illegitimate.
• Instead, it would allow the culture and infrastructure to catch up with the technology

The complete “walk before you run” sequence

The overall argument can therefore be represented as:

1. Technology arrives
↓
2. Artists experiment with it
↓
3. Artists develop mastery
↓
4. A community forms
↓
5. New aesthetics emerge
↓
6. A cultural identity develops
↓
7. Standards and forms of quality develop
↓
8. Legal and institutional norms develop
↓
9. Distribution systems adapt
↓
10. Commercial integration occurs

Skipping it creates the problems now being experienced and i'm expecting even harsher and stricter measures taken against AI generated content in the near future

And therefore:

>The “walk before you run” argument: the concern isn't simply that there is too much AI music. It's that AI music skipped the cultural maturation process that normally transforms a new technology into a respected artistic medium.

Yet many here are DEMANDING THAT RESPECT INSTEAD OF EARNING IT

AND RIGHT NOW THE GRIFTERS/SPAMMERS/SCAMMERS are the most visible members of our community to the wider public.

UMG vs. DistroKid: The AI music distribution fight is getting real.

Universal Music Group has filed a federal lawsuit against DistroKid, one of the biggest independent music distributors, accusing it of deceptive trade practices and copyright infringement.

This is worth paying attention to if you use Suno and distribute your music through DistroKid.

What happened?

On September 15, 2026, UMG and its Capitol Records operations filed a lawsuit in the U.S. District Court for the District of Delaware.

The lawsuit alleges that DistroKid has been distributing large amounts of AI-generated music and unauthorized versions of copyrighted recordings across streaming platforms.

UMG calls this an “AI-slop pipeline.”

But here's the important part:

UMG is not claiming that all AI-generated music is automatically illegal.

According to Music Business Worldwide's reporting on the complaint, UMG specifically distinguishes between AI music that is clearly disclosed and allegedly infringing or deceptive content.

The core allegations include:

• DistroKid allegedly distributes mass-generated AI content that crowds out legitimate artists.
• Unauthorized remixes, altered recordings, and copyrighted music are allegedly being distributed through the platform.
• DistroKid allegedly gives the impression that its catalog consists of artist-backed releases created and owned by real human artists.
• UMG alleges that infringing and deceptive uploads can divert streams and revenue from legitimate rights holders.

These are allegations, not findings of fact. DistroKid's response was not available in the initial reports I found.

In one example, the lawsuit sites an account that uploaded 4000+ "Low fi beat tracks" within a year. That's more than 11 a day!

Absurd! And IMO It's wrong of Distrokid to facilitate that.

What I find interesting

UMG is going after the distributor in this case, not just individual uploaders.

That could put pressure on the entire DIY distribution model.

If distributors become more cautious about AI music, we could see:

• More AI disclosure requirements.
• More content screening before release.
• More scrutiny of large catalogs and mass uploads.
• More disputes over who owns the rights to a track.
• Potential changes to how AI-generated music is distributed and monetized.

My questions for the Suno community

1. Are you currently distributing Suno music through DistroKid?
2. Have you experienced any AI-related rejections, takedowns, or copyright claims?
3. Do you think distributors should require AI disclosure?
4. Should fully AI-generated music be treated differently from music that uses AI-assisted vocals, instruments, or production?
5. Does this make you more cautious about releasing Suno music commercially?

Sources:

• Music Week — Universal Music Group sues DistroKid over alleged copyright infringement (September 15, 2026)
• Music Business Worldwide — Universal Music Group sues DistroKid, accusing it of “unlawful practices” and “flooding platforms with AI-generated slop” (September 15, 2026)
• Law360 — UMG Says DistroKid Is Copyright-Infringing “AI-Slop Pipeline” (September 15, 2026)

“Long way to grow” was written 2 decades ago at the age of 19.

I had always thought of it as “Piano song” since I wrote the core of it on the keyboard back then.

But the only demo I had of it, was this one, where I play the core of it on the guitar.

I think the song was written about all the expectations we have growing up and all these grand ideas Of how things will be only to find yourself Up creek and without paddle.

My friend Scarecrow at the time would often bust my balls to do more and be more, often bringing an attitude over when he hung out. I remember writing the lines “See I don’t mind, I’ll take my time. So check your pride at the door! If you’ve somewhere better than here, what are you waiting for?” About this dynamic between us at the time.

The original demo’s lead vocals are still in lead, Witchdoctor backing me up on the vocals, My guitar
Is driving the train and the witch doctor adds some little quips throughout.

The past couple days I have added Piano (classical and honky tonk stye), Tuba, Trombone, Choir and a very subtle cello in Bandlab with Midi.

With Suno, I added a Bass and Percussion. There technically is a very very subtle backup vocalist from suno but it’s so faint you can’t really hear it in the mix.

Visually you are seeing on the TV, an avatar of me from my late teens that has been animated by me using Motion Sync in OpenAI. Basically I’m puppetting the avatar with my real movements and guitar playing.

This is actually the second time posting this composition. The original post leaned more into Suno Performance, this one leans into the original.

Penny for ye thoughts?

“A song” was written nearly 2 decades ago at the age of 20. It is just a short instrumental guitar part that I wrote and I believe it was recorded simply on my laptop’s internal mic, which obviously left the quality in a
Limited state.

But it’s the only recording I have of it.

“A song” as a name was in reference to its key, in this case the key of A.

And it was later absorbed by a larger composition serving as the bridge.

That larger composition, never was completed (never wrote the lyrics)

I may try to later in the project.

But I figured I’d share it in this form, now.

All of the other instruments were played by me with Midi in bandlab.

And include, Cello, Violin, Male Choir, Female Choir.

No suno was used in this one, but I still am sharing it in this community cause, well this is my community.

Visually you are seeing on the TV a clip I took of the moon from my sun room one night
Using my Supertelephoto lens. I made the mistake of touching the camera during the shot
With such a long lens causing the slight jitter, sigh.

Penny for ye thoughts?

TL;DR: The DOJ is backing the argument that AI training on copyrighted material can be fair use—but it did not rule that AI training is automatically legal. This is a major development for AI companies like Suno, but the courts still have the final say.

There’s been some confusion about the recent DOJ news, so here’s what actually happened:

On September 1, 2026, the U.S. Department of Justice filed a Statement of Interest in the ongoing New York Times v. OpenAI/Microsoft litigation.

The DOJ is arguing that using copyrighted material to train AI models can qualify as fair use, describing AI/LLM training as highly transformative and arguing that the benefits of AI development can outweigh the competitive harm claimed by copyright holders.

The DOJ also argues that requiring AI companies to obtain licenses for all copyrighted material used in training could have significant consequences for U.S. AI development, innovation, and national security.

But here's the important part:

🚨 The DOJ did NOT issue a ruling saying that all AI training is fair use.

It is a legal filing asking the federal court to adopt that interpretation. The judge still has to decide the case. The DOJ's filing is influential, but it is not itself binding law or a court judgment.

Why does this matter to Suno?

Suno's situation involves music, recordings, compositions, and copyright, whereas the particular DOJ filing concerns litigation involving written works and AI language-model training.

>What we can accurately say is that the U.S. government has now formally taken the position in federal court that AI training can constitute fair use of copyrighted material and that overly broad copyright restrictions on AI training could harm U.S. innovation.

The eventual court decisions in the various AI copyright cases could have major implications for companies like Suno — and for creators whose work is used to train AI.

For now, the legal question is far from completely settled.

Do you think the DOJ is correct in their recommendation?

Or do you think they are cutting corners on long established IP rights in order to "win" the AI race with China?

Just thought this could be an interesting experiment

How well does suno faithfully "cover" your original song when you want it to?

A side by side comparison of 2 sections of one of my basement tapes vs a fully Suno performed cover of the same composition.

The focus is on guitar, vocal, melody and composition.

TL;DR: Jason Isbell and other musicians are proposing a class action lawsuit against Suno, arguing that the AI music generator can exploit artists’ identities—not just copy copyrighted songs—by generating music that allegedly reproduces recognizable voice and stylistic traits when users invoke their names. Suno says it generates original music and blocks artist-specific prompts. The big legal question: does imitating an artist’s style cross the line from permissible stylistic influence into unlawful appropriation?

The news

Jason Isbell, David Lowery of Camper Van Beethoven/Cracker, Guy Forsyth, and saxophonist Eduardo “Ed” Calle have filed a proposed class-action lawsuit against AI music generator Suno in Massachusetts federal court.

The case is Lowery et al. v. Suno, Inc., No. 1:26-cv-14005, filed August 31, 2026.

This one is notably different from the copyright cases already facing AI music companies. The musicians do not plead a copyright-infringement count in this complaint. Instead, they allege that Suno unlawfully exploits musicians' identities through state right-of-publicity laws and related theories. The complaint also includes an Illinois biometric-privacy claim concerning alleged “voiceprints.”

The plaintiffs' theory is essentially that Suno has allegedly encoded performer-specific information so that an artist's name can operate as a kind of “retrieval key” for outputs associated with that performer—potentially including music, descriptions, vocals and imagery. That is the plaintiffs' allegation, not something a court has yet found Suno actually does.

What the complaint says happened

One of the central examples involves Jason Isbell. According to the complaint, prompting Suno with his name generated an Americana song titled “Paper Bell.” The plaintiffs characterize its vocals as having Isbell-associated traits such as clear male vocals and a country twang; Suno reportedly described the result as contemporary Americana with fingerpicked acoustic guitar. The complaint does not claim that “Paper Bell” copied one particular Isbell song.

The lawsuit also attacks Suno's artist-name safeguards.

Suno has publicly said that it does not allow prompts for specific artists or copyrighted songs and that, when an artist is referenced, the artist's name is removed and the request is redirected toward descriptive musical characteristics.

The plaintiffs say those restrictions can be circumvented.

Suno's response

Suno says the claims lack merit and says it will defend itself.

Its public response is that Suno exists to help people create new, original music, rather than trade on artists' names, and that it maintains protections including blocking prompts for specific artists and copyrighted songs.

So at this point we have sharply competing factual and legal claims. Suno has not been found liable in this case.

The interesting legal question isn't really “Can you copyright a style?”

This is where I think the case gets genuinely interesting.

Under U.S. copyright law, an artistic style by itself generally isn't protected as a separate copyrighted element.

The U.S. Copyright Office addressed this exact AI issue in its 2024 Digital Replicas report, stating that copyright law does not protect artistic style as a separate element of a work. An “in the style of” output can still infringe copyright if it reproduces protectable expression from a particular work—but similarity of style alone is a different issue.

So an artist generally cannot say:

>

That would collide with a basic principle of copyright: later artists remain free to learn from genres, techniques, ideas and other unprotectable building blocks.

But right of publicity is a different body of law.

It asks whether someone commercially appropriated an identifiable person's identity, which can include things such as a person's name, likeness and—in some jurisdictions—voice or other distinctive attributes.

And that's where the line becomes fuzzy.

Why the plaintiffs have a serious argument

The closest famous precedent is Midler v. Ford Motor Co. from the Ninth Circuit.

Bette Midler refused to sing for a Ford commercial. The advertisers then hired one of her former backup singers and deliberately instructed her to imitate Midler's distinctive voice.

Importantly, Ford had licensed the underlying song. So copyright ownership wasn't the decisive issue.

The Ninth Circuit held that deliberately imitating the distinctive and widely known voice of a professional singer to sell a product could constitute appropriation of that singer's identity under California law. But the court deliberately kept its holding narrow; it did not announce that every vocal imitation or stylistic similarity is unlawful.

The Suno plaintiffs are effectively pushing that identity concept into generative AI.

Their strongest version of the argument isn't:

“Jason Isbell owns Americana.”

It's closer to:

“A commercial system allegedly recognizes Jason Isbell specifically by name and then generates a package of performer-specific vocal traits, musical characteristics and associated imagery because users asked for Jason Isbell.”

If a court accepts that characterization, the case starts looking much more like alleged exploitation of identity than ordinary musical influence.

Why Suno also has a serious legal argument

The other side is that there is a huge difference between reproducing someone's actual identity and producing something merely reminiscent of their style.

And existing law is not uniform.

The Copyright Office specifically noted this tension. It observed that some right-of-publicity cases have treated aspects of a person's performing style as potentially part of their persona, while other courts have warned against using publicity rights to prohibit the portrayal of someone's personality or performance style.

The Copyright Office ultimately did not recommend creating a federal property right in artistic style itself, in part because doing so could interfere with later creators' ability to build on existing artistic traditions.

That matters far beyond AI.

Human musicians imitate influences constantly. Blues, hip-hop, punk, country, jazz and pop all evolve through musicians absorbing recognizable techniques from earlier artists.

A rule saying “this output sounds recognizably like Artist X, therefore Artist X's publicity rights were violated” could potentially reach much further than literal voice cloning.

The court will therefore have to confront a very difficult boundary:

When is an AI merely generating unprotectable stylistic characteristics, and when is a commercial system exploiting enough performer-specific attributes that it is effectively selling access to that person's identity?

The big takeaway

I don't think this case can accurately be reduced to “artists are trying to copyright their style.”

They aren't asserting copyright ownership over style in this complaint.

They're attempting something different: using right-of-publicity and identity laws to argue that a commercial AI service cannot allegedly turn a musician's name plus recognizable personal attributes into an on-demand product without permission.

Whether that theory succeeds is very much an open question.

And whatever courts decide could have consequences well beyond Suno—not just for AI music generators, but for the legal boundary between inspiration, imitation, impersonation and identity.

Where would you draw that line?

NO SUNO OR AI WAS USED IN THIS SONG AT ALL.

100% “traditionally” produced.

I’m doing so not because I’ve changed my mind on the utility of AI, far from it.

But simply because this song already felt pretty complete, unlike most of my old demos which are much more bare relative to how I’d like them to be.

And yes, I wanted to prove a point to the Anti’s.

“Stink eye” was written 2 decades ago at the age of 19.

There is not much meaning to the song other than the obvious.

Just a 2 chord jammer.

The one caveat I’d say is my personal reality did not match the bravado of the song lol.

One summer evening with my Mom out of town the Witchdoctor and I laid out the base track
Lead vocal, Rhythm Guitar (Me) and WD soloing on a 2nd acoustic.

Not only is it Take 1, it was the first time he and I played it all the way through together.

Then we took that recording over to Calico Jack’s home studio in his parent’s basement

and he added the bass, which honestly makes the song, and drums, although that may have been his brother, I can’t remember.

And then we had Scarecrow come by and sing vocals on it, in tandem with me.

Today I added those slightly higher octave vocals in the 2nd chorus, and was surprised how little my voice has changed, if at all.

Then I added, Tuba, and Organ and Choir in midi with Bandlab.

Visually, your seeing a 19 year old me avatar/dopppleganger (clothes and all) from the very first time I ever performed Live in front of a crowd (coffeeshop)

I animated the avatar using motion sync in Cannonstudio. So the body language is actually me. Recorded that today in my sun room in front of A white screen. Did pretty good overall, but it flattened all the nuance in my individual fingers on the guitar playing so that left something to be desired there.

I still have that sound recording, and I have another live performance video of us playing this song live with a full band.

That one is cool cause we really open it up at the end, unlike this recording that remains somewhat subdued.

Penny for ye thoughts!?

Thought exercise.

Which Scenario results in the more "Human" music?

Especially interested in hearing Auntie opinions

A. AI generates a composition including lyrics that's great, then a human band performs it faithfully to the AI output.

B. Human creates a composition, including lyrics, structure, melody, pacing, etc, then an AI executes the vision faithfully.

This question get's at the heart of not just human vs AI music but what it means to be a human.

Are we more our physical presence or are we more our consciousness? Obviously it's both, but I'm curious in the context of music, if you had to choose between A or B, which is more human?

“A bad song” was written nearly 2 decades ago in college. 

This is the song referenced in my other post from today "At what point is it considered AI music?"
https://www.reddit.com/r/Suno/comments/1w2se09/at_what_point_is_it_considered_ai_music/

It’s a sarcastic social commentary I made of American Culture at the time but I think probably more true now then even then. 

The original demo is still driving the mix. My original Rhythm guitar and a the Witchdoctor on Djembe.

My original lead vocals are still the lead vocals you are hearing. (For better or worse) with Witchdoctor singing some backups.

Recently, I brought it into Bandlab and add midi piano, french horn, trombone, Tuba and Choir and probably forgetting some. 

With Suno, I added a bass, percussion, and a subtle backup vocal.

The visuals are a real open mic performance of us at college performing this song, although your not hearing this particular performance
I simply did my best to sync up the visual with the recorded audio. I am on the right and Witchdoctor on the left. 

Penny for ye thoughts?

Rather than passive aggressively downvote, or leave snide comments, if you have an arguement (not personal attacks) you'd like a genuine reply to, I'm willing to oblige. As a traditional musician + composer, who see's value in AI, I think I'm in a good position to do so .

"Good show, me" , written and composed at the age of 19 two decades ago. I wrote and recorded the bare demo at the time, and recorded it on my mini dv camera which is literally the clip on the tv. I wrote and recorded it within an hour as a house party was happening downstairs hence the background talking you can hear at the start and end.

The guitar and lead vocal are from the original recording. It did have me attempting to whistle the melody in the instrumental section but I suck at whistling and I removed it from this new mix lol.

In Band lab with my midi keys I added, piano, tuba, choir, french horn, and put the whistle back in. I really enjoy layering instruments with midi to create a subtle "Symphony" effect.

With Suno I added a very subtle backing vocal, Bass and percussion.

Not one of my more memorable songs but I still like it.

Anyhow, penny for ye thoughts?

TL;DR: ARIA is drawing a line between using AI as a tool and letting AI be the artist.

Starting with the ARIA chart dated August 31, 2026, recordings that are wholly generated by AI will no longer be eligible for Australia's official music charts. Fully AI-generated works will also be ineligible for ARIA Awards.

AI-assisted music is still allowed, but there are conditions: the recording must be substantially human-made, with humans responsible for the core creative performance, and any generative AI used must be authorized and lawful.

The move follows controversy over an AI-involved version of Madonna's Like a Prayer by Australian producer Josh Fawaz. The track became a major hit, spending 16 weeks in Australia's Top 20, before Fawaz disclosed that AI had been used for elements including the vocals and drums.

ARIA's decision also aligns with new principles from the International Federation of the Phonographic Industry (IFPI), which are pushing for chart-eligible music to remain substantially human-made and for AI-involved recordings to be developed lawfully and without manipulation.

So the policy isn't really an outright “AI music ban.”

The distinction is:

• ❌ Wholly AI-generated track: Not eligible
• ❌ Fully synthetic music competing for ARIA Awards: Not eligible
• ✅ Human artists using AI as part of their creative process: Potentially eligible
• ✅ AI-assisted music that remains substantially human-made: Eligible, subject to the rules

The bigger question is whether this is the beginning of a broader industry standard. As AI-generated music becomes increasingly convincing, charts and awards are going to have to decide what exactly counts as a human performance — and where using a tool ends and replacing the artist begins.

Sources:

• ARIA / Australian reporting: ARIA's new chart eligibility rules and the Josh Fawaz controversy
• IFPI: Global principles for AI and official music-chart eligibility
• Associated Press: Reporting on the ARIA policy and AI-generated music's eligibility for awards

What do you think? Should fully AI-generated songs be allowed to compete on the same charts as human artists, or is ARIA making the right call?

"Last time I did that" was written when I was 21 years old, nearly 2 decades ago.

This is the second version of the song I have shared. The first version had a Suno Lead singer with me as backup, and this one is my original lead vocal and a touch more midi Tuba work :)

To me, this song is about that feeling you get when you've been moping around in lethargy for too long and finally something defiant suddenly swells up inside you and launches you into action. This was a regular song I performed at open mics back in college.

The original demo is still driving the mix. My original Rhythm and Lead Guitars (multi tracked) is all you are hearing in terms of guitar.

My original lead vocals are still the lead vocals you are hearing. (For better or worse)

Additionally, I had a buddy record this demo with me back in the day and he was on backup vocals and Djembe. Add to that, a few sporadic backup vocals from a separate (but from the time) demo where me and 2 of my friends got pretty drunk and let loose. Maybe someday, I can release that "drunk demo" as I always called it as another version.

Recently, I brought it into Bandlab and add midi Keys, Cello, Tuba and Choir.

With Suno, I added a A brass section, a bassist, keyboard , kick drum, and a back up vocalist very subtly mixed in.

The visuals are a real open mic performance of us at college performing this song, although your not hearing this particular performance
I simply did my best to sync up the visual with the recorded audio. I am on the right and Witchdoctor on the left.

Penny for ye thoughts?

"Directions" was composed decades ago at the age of 18.

It was part of the very first 10 or so songs I’d ever written.

It was meant to be a funny catchy little bop about obvious matters of
A young man.

The original demo (recorded at age 20) is still very much in the driver's seat and
includes ALL the multi-tracked guitar parts and the lead vocals (which I also sang
As my own backup on)

Back then I was so insecure about the sound of my voice that I decided double layering my voice somehow made it better, lol not sure what sense that made.

In Bandlab Midi I added, well probably too much lol but, Cello, Tuba, French Horn, Organ, Choir, Snare drum.

With Suno, I added a back up vocalist rotoscoped to the original, but left my original vocal as the lead vocal. A bass that is rotoscoped to a low end guitar line I had multi tracked back then, a Djembe, and some Trumpet hits here and there.

The visuals are a slightly modified face of a real photo of me from the age of 18 around the time I wrote the song and I used a Motion track to control the image like a puppet. So the body language of the avatar is actually me. It did a crappy job of translating my guitar playing. Definitely could be better but I ran out of credits trying to dial it in. Then I brought it all together in Adobe Premiere. The Bob Dylan postergeist was not planned but I didn’t have the credits to fix it even if I wanted to lol.

At the end where I laugh and say “Probably sounds like shit” that’s from the original demo.

This is actually the second version of the song I’ve posted. The first was the very first video I put on my channel and was 100% performed by suno (but obviously faithfully covered to this one) To be honest, I still prefer that All Suno one. I’m remaking it now because the scope of my project changed from originally supposed to be all Suno performed to this Hybrid version I am now focused on.

You can see the SUNO ONLY Performance of this composition here
https://www.youtube.com/watch?v=bdFKBAfP3gw

If you happen to listen to both, I'd be super curious to hear which you prefer, if you are like me and prefer the SUNO only or if you prefer the hybrid?

Penny for ye thoughts?