Tag: AI and music

@ArtistRights Symposium 5 Panel: Beware the AI Opt-In: Hidden Right of Publicity Waivers

Registration is open for Artist Rights Symposium 5 at Eventbrite!

The Artist Rights Symposium 5 is coming on October 19 at Tweed Recording in Athens, GA hosted by the University of Georgia Terry College of Business Music Certificate Program and the School of Policy and International Affairs, and the Artist Rights Institute. More about the Symposium here and register on Eventbrite here.

As AI companies like Suno and Udio settle copyright lawsuits with record companies and music publishers, they introduce new licensing arrangements. Artists and songwriters may be offered the “opportunity” to “opt in” to the settlements and “new revenue streams”. But not all revenue streams are equal, remember streaming? Remember “Let’s help that plucky startup Spotify”? We have seen this movie before and we know how it ends.

Why a Suno Copyright Settlement Is Different from Spotify, LimeWire—or Even Bartz

The music industry has settled copyright infringement disputes before. Spotify’s mechanical royalty settlements in David Lowery’s class action resolved claims over unpaid or improperly licensed compositions. LimeWire’s settlement addressed unauthorized reproduction and distribution of copyrighted recordings. In both cases, the essential bargain concerned compensation for past copyright infringement, not permission to reproduce the identity of the human beings who created the music. Identity theft never came up.

Suno presents a fundamentally different problem. Its generative AI technology can potentially reproduce not merely copyrighted expression but recognizable characteristics of an artist’s voice, performance and identity—what Chris has called  personhood signals. A settlement that licenses recordings for future training may therefore implicate rights that neither the record company nor publisher controls. That’s why they need the “opt in” agreement to waive these rights because the AI platform doesn’t want to pay for the artist’s identify if the artist would even consider that deal. Many artists and songwriters, if not most, will not.

The Bartz v. Anthropic settlement illustrates the limitation of treating AI disputes exclusively as copyright cases. Bartzcompensates copyright owners for specified historical conduct involving pirated books. It settles the claims brought in litigation, so does not expressly resolve the separate problem of extracting and commercializing human identity. That is what a conventional copyright settlement misses.

Settling infringement of the work does not settle exploitation of the person. And no copyright settlement should quietly become consent to commercialize someone’s voice, identity or personhood.

So artists, their managers and lawyers are asking, what exactly are we being asked to agree to this time? Before you even get to the “new revenue stream,” an opt-in for an apparent copyright settlement may also include permission for future AI training, broad releases, perpetual covenants not to sue, or waivers of an artist’s rights in their name, voice and identity. These are fundamentally different transactions, with potentially permanent consequences. Before signing, creators need to understand not only what they will be paid, but what rights they are surrendering, how those rights may be exploited, and whether they can ever get them back.

Meet the Panelists

Chris Castle will lead the closing panel at Artist Rights Symposium 5, that warns “Beware the AI Opt-In: Hidden Right of Publicity Waivers.” The panel brings together four speakers with complementary perspectives on artist advocacy, litigation and music-industry licensing.

Jen Jacobsen, Artist Rights Alliance.  Friend of the show Jen Jacobsen is a long time artist rights advocate who brings the artist-advocacy perspective, focusing on meaningful consent, transparency and protecting musicians from agreements that may surrender irreversible grants of rights with hidden waivers.

George Zelcs of Korein Tillery.  George Zelcs is the lead counsel on the Jason Isbell/David Lowery personhood class action against Suno to help us understand the implications of “opt in” agreements with waivers on rights of publicity and the practical question of what remedies artists should retain.

Gwendolyn Seale — Gwen Seale PLLC.  Gwen brings experience representing songwriters and music copyright owners, with particular relevance to licensing authority, co-writer and co-owner interests, compensation and contractual protections.

Tim Kappel — Wells Kappel.  Tim is a copyright expert and artist rights advocate who successfully argued in federal trial court and the Fifth Circuit Court of Appeals that his client Cyril Vetter should retain foreign rights as part of a statutory copyright termination— a case now before the U.S. Supreme Court. 

The panel is scheduled for 3:00–4:00 p.m. on October 19. 

Three Deals Disguised as One?

An AI opt-in may combine three distinct transactions as part of the copyright settlement and go-forward license: compensation for past unauthorized uses (probably not shared by the label, but check), licensing for future model training, and revenue sharing from products built on those models (but don’t hold your breath). But accepting settlement compensation for yesterday’s conduct should not automatically mean consenting to tomorrow’s exploitation. 

There may also be a fourth transaction: authorization to use an artist’s name, voice, likeness or other personhood rights which could not have been brought by a record company or music publisher because they do not control those rights.

A copyright license is not permission to commercialize personhood.

The panel will examine how these rights can become entangled, particularly when record companies, publishers or other representatives may not control all the rights an AI company seeks.

What Happens After You Sign?

The panel will address the questions that can be overlooked when attention focuses on the size of a settlement or the promise of future royalties.

What exactly is inside the AI company’s “walled garden”? How are revenues attributed to particular recordings or compositions? Can users generate synthetic voices or prompt an artist’s name? Can artists examine the product before consenting? 

And what happens when an artist wants out later when the term expires (if there is a term)? Deleting a source recording does not necessarily remove its influence from an already-trained model. Termination, machine unlearning, continuing model exploitation and rights that survive expiration may determine whether an opt-in becomes a one-way door in an AI Hotel California.

Register for Artist Rights Symposium V on Eventbrite

Related reading: The AI “License” That Isn’t: A Musician’s Checklist for Spotting a Covenant Not to Sue Censorship Trap in Disguise.

Infrastructure, Not Aspiration: Why Permissioned AI Begins With a Hard Reset

Paul Sinclair’s framing of generative music AI as a choice between “open studios” and permissioned systems makes a basic category mistake. Consent is not a creative philosophy or a branding position. It is a systems constraint. You cannot “prefer” consent into existence. A permissioned system either enforces authorization at the level where machine learning actually occurs—or it does not exist at all.

That distinction matters not only for artists, but for the long-term viability of AI companies themselves. Platforms built on unresolved legal exposure may scale quickly, but they do so on borrowed time. Systems built on enforceable consent may grow more slowly at first, but they compound durability, defensibility, and investor confidence over time. Legality is not friction. It is infrastructure. It’s a real “eat your vegetables” moment.

The Great Reset

Before any discussion of opt-in, licensing, or future governance, one prerequisite must be stated plainly: a true permissioned system requires a hard reset of the model itself. A model trained on unlicensed material cannot be transformed into a consent-based system through policy changes, interface controls, or aspirational language. Once unauthorized material is ingested and used for training, it becomes inseparable from the trained model. There is no technical “undo” button.

The debate is often framed as openness versus restriction, innovation versus control. That framing misses the point. The real divide is whether a system is built to respect authorization where machine learning actually happens. A permissioned system cannot be layered on top of models trained without permission, nor can it be achieved by declaring legacy models “deprecated.” Machine learning systems do not forget unless they are reset. The purpose of a trained model is remembering—preserving statistical patterns learned from its data—not forgetting. Models persist, shape downstream outputs, and retain economic value long after they are removed from public view. Administrative terminology is not remediation.

Recent industry language about future “licensed models” implicitly concedes this reality. If a platform intends to operate on a consent basis, the logical consequence is unavoidable: permissioned AI begins with scrapping the contaminated model and rebuilding from zero using authorized data only.

Why “Untraining” Does Not Solve the Problem

Some argue that problematic material can simply be removed from an existing model through “untraining.” In practice, this is not a reliable solution. Modern machine-learning systems do not store discrete copies of works; they encode diffuse statistical relationships across millions or billions of parameters. Once learned, those relationships cannot be surgically excised with confidence. It’s not Harry Potter’s Pensieve.

Even where partial removal techniques exist, they are typically approximate, difficult to verify, and dependent on assumptions about how information is represented internally. A model may appear compliant while still reflecting patterns derived from unauthorized data. For systems claiming to operate on affirmative permission, approximation is not enough. If consent is foundational, the only defensible approach is reconstruction from a clean, authorized corpus.

The Structural Requirements of Consent

Once a genuine reset occurs, the technical requirements of a permissioned system become unavoidable.

Authorized training corpus. Every recording, composition, and performance used for training must be included through affirmative permission. If unauthorized works remain, the model remains non-consensual.

Provenance at the work level. Each training input must be traceable to specific authorized recordings and compositions with auditable metadata identifying the scope of permission.

Enforceable consent, including withdrawal. Authorization must allow meaningful limits and revocation, with systems capable of responding in ways that materially affect training and outputs.

Segregation of licensed and unlicensed data. Permissioned systems require strict internal separation to prevent contamination through shared embeddings or cross-trained models.

Transparency and auditability. Permission claims must be supported by documentation capable of independent verification. Transparency here is engineering documentation, not marketing copy.

These are not policy preferences. They are practical consequences of a consent-based architecture.

The Economic Reality—and Upside—of Reset

Rebuilding models from scratch is expensive. Curating authorized data, retraining systems, implementing provenance, and maintaining compliance infrastructure all require significant investment. Not every actor will be able—or willing—to bear that cost. But that burden is not an argument against permission. It is the price of admission.

Crucially, that cost is also largely non-recurring. A platform that undertakes a true reset creates something scarce in the current AI market: a verifiably permissioned model with reduced litigation risk, clearer regulatory posture, and greater long-term defensibility. Over time, such systems are more likely to attract durable partnerships, survive scrutiny, and justify sustained valuation.

Throughout technological history, companies that rebuilt to comply with emerging legal standards ultimately outperformed those that tried to outrun them. Permissioned AI follows the same pattern. What looks expensive in the short term often proves cheaper than compounding legal uncertainty.

Architecture, Not Branding

This is why distinctions between “walled garden,” “opt-in,” or other permission-based labels tend to collapse under technical scrutiny. Whatever the terminology, a system grounded in authorization must satisfy the same engineering conditions—and must begin with the same reset. Branding may vary; infrastructure does not.

Permissioned AI is possible. But it is reconstructive, not incremental. It requires acknowledging that past models are incompatible with future claims of consent. It requires making the difficult choice to start over.

The irony is that legality is not the enemy of scale—it is the only path to scale that survives. Permission is not aspiration. It is architecture.