The NO FAKES Act cleared the Senate Judiciary Committee this week. Unanimous voice vote, on its third attempt, the first time it's made it out of committee.

It now heads to a full Senate floor vote.

Before you file this under "celebrity protection bill," stop. That's not what this is.

Right now, there is no federal law protecting your voice. Not for the session vocalist in East Nashville. Not for the Tejano singer in San Antonio. Not for the Indian classical musician whose recordings are sparse and whose estate is informal.

Any of them can have their voice cloned, packaged, and monetized with no uniform federal cause of action to stop it. The NO FAKES Act changes that.

In this episode:

What the bill actually creates: a federal property right over voice and likeness, for every American. For the first time!

The enforcement number that matters: $750,000 per work for platforms that fail to comply with notice and takedown.

Tthe figure that forces Spotify, YouTube, TikTok, and AI tool companies to build real infrastructure instead of playing whack-a-mole. The 70-year post-mortem right most outlets aren't covering, and why it's a cultural preservation question, not a technicality.

What happens next?

Full Senate vote, then the House, then a signature — and why the RIAA's "later this year" timeline reads as cautiously real. 

Senator Marsha Blackburn of Tennessee as lead sponsor, and why Music Row has been the loudest constituency pushing this forward

The political will exists right now in a way it simply didn't in 2024 or 2025. That can change quickly in either direction.

If you're a session musician, independent artist, or estate representative without a federal cause of action today, this bill is the first one that closes that gap. Follow its progress.

I'm Raia. This is The Co-Write Room. Daily at the intersection of AI, music, and the business reshaping both.

Follow and rate the show wherever you listen.

• (00:00) - Cold Open

What if the song that went viral last week was purchased? And what if the royalties you didn't earn from it were stolen on top of that? In Episode 9, Raia Kumar tracks four stories — manufactured virality, the first federal AI streaming fraud conviction, a billion-dollar funding round mid-lawsuit, and a landmark bill that could finally give independent songwriters a legal seat at the table.

In this episode:

• Labels are paying regular people around forty dollars per one hundred views to create short-form videos using specific tracks — manufacturing what the algorithm reads as organic momentum. Trevor Noah said it out loud. The industry has known for years.

• Michael Smith, 54, of North Carolina pleaded guilty on March 19, 2026 to using AI to generate hundreds of thousands of songs and bots to fake billions of streams — collecting over ten million dollars in royalties that belonged to working musicians. He is the first person in American history convicted of a federal crime for AI-assisted music streaming fraud. The key legal distinction: the crime wasn't AI music. The crime was manufacturing the audience.

• Suno closed a four hundred million dollar funding round this week — valuation five point four billion dollars, more than double seven months ago — while actively being sued by major labels for copyright infringement. Their new AI model is "built in partnership with the music industry," which turns out to mean a deal with Warner Music Group. One major label. Not independent songwriters.

• Within 72 hours of each other, Suno and Udio filed nearly identical motions to seal the same piece of information in their separate federal cases: the total number of audio files each company used to train their AI. The argument is trade secret protection. The effect, if granted, is that the most important number in determining the scope of copyright damages may be permanently hidden.

• On June 1, the Artists Rights Alliance launched a national campaign behind the Protect Working Musicians Act, reintroduced by Representative Deborah Ross of North Carolina. The bill would give independent musicians an antitrust exemption — the legal right to collectively negotiate with AI companies and streaming platforms. That right does not currently exist.

• What is your organization's position on the Protect Working Musicians Act, and are you in contact with Representative Ross's office?

  0:00 | Cold Open | The purchased song. The stolen royalties. The thesis.
  0:30 | Welcome | Raia opens the episode. June 4. Four stories, one thread.
  0:50 | Story 1: Labels Buy Virality | Trevor Noah's clip. Labels paying ~$40/100 views for short-form videos. What that means for independent artists competing against marketing spend.
  2:15 | Story 2: The First Conviction | Michael Smith guilty plea. AI-generated songs + streaming bots. $10M stolen. The legal distinction that defines what comes next.
  3:45 | Story 3: The Money & the Cover | Suno's $400M raise at $5.4B. The Warner Music Group deal. Suno and Udio's near-identical motions to seal training data volume in federal court.
  5:15 | Story 4: The Response | Artists Rights Alliance launches Protect Working Musicians Act campaign. Rep. Deborah Ross. What the bill does. The CTA for independent artists.
  6:15 | Closing | The frameworks being built now are being built without independent creators at the table.

• (00:00) - Cold Open

The TAKE IT DOWN Act is now law. Which means platform enforcement rules are in effect — and there are two things buried in the fine print that matter far more than the headline.

In this Newsflash, Raia breaks down:

• What the law actually requires covered platforms to do — and the 48-hour federal deadline that is now non-negotiable
• Why known identical copies must also be removed, closing a loophole bad actors have exploited for years
• The legal distinction that changes everything: consent to create does not equal consent to distribute
• What this law does not do — and why the burden of discovery still falls on you

For musicians creators, AI voice cloning and visual likeness tools have made it trivially easy to generate convincing content featuring real artists — posted without permission, without compensation, and until now, without any clear path to removal. That just changed.

You should know where to find each platform's reporting tool. The 48-hour clock doesn't start until you trigger it.

The Co-Write Room is an AI-powered podcast covering the intersection of artificial intelligence, music, and creative rights.

• (00:00) - Cold Open

A new study proved that labeling music as AI, even when it's human-made, causes listeners to emotionally disconnect. The label changed. The music didn't. 

This week, Raia covers the infrastructure being built to apply those labels: 

• The Suno lawsuit expanding to 61,000 recordings
• Apple Music's AI model fingerprinting
• Quicksilver browser extension
• Protect Working Musicians Act
• Johnny Cash's estate suing Coca-Cola under Tennessee's ELVIS Act
• YouTube's automatic AI detection rollout. 

The infrastructure of proof is being built right now. So is the infrastructure of protection. They are not moving at the same speed. And they are not being built for the same people.

In this episode:
The Suno lawsuit expansion: two weeks of Audible Magic audio fingerprinting surfaced 61,026 recordings and the labels called it a small fraction of total infringement. What the technology just proved is possible in federal court.

What Universal and Sony won't tell you: they're fighting for master recordings, not the underlying compositions many of which belong to independent Nashville publishers and songwriters with no seat in that Massachusetts courtroom

Apple Music's internal count: one third of all submissions are now AI-generated. Less than 0.05% of listening time goes to those tracks but in a pro-rata royalty pool, AI tracks don't need listeners to do damage. They just need to exist. Apple's proprietary technology that can identify not just whether a track is AI-generated, but which AI model produced it and what that means for mandatory disclosure at the distribution level

Quicksilver: the browser extension from the University of Chicago team behind Glaze and Nightshade. Press Analyze while a song streams. It scans for inaudible AI audio artifacts on your device. Nothing uploaded. And the research behind it puts nearly 50% of weekly new music releases as AI-generated

The Protect Working Musicians Act, reintroduced May 21st by Rep. Deborah Ross with Tennessee's Rep. Steve Cohen as co-sponsor. Why the antitrust exemption framing matters and why naming AI developers as negotiating counterparties is a significant shift

Nashville: the Johnny Cash estate suing Coca-Cola in federal court under the ELVIS Act. Why Coca-Cola's soundalike defense, if it succeeds, becomes a roadmap for AI voice replication. 

YouTube's automatic AI content detection rollout and the gap between who gets likeness protection (signed artists at CAA, UTA, WME, major management) and who gets caught by the enforcement system (everyone)

Look up the Protect Working Musicians Act. Find your representative and contact them. The Artist Rights Alliance has a one-click contact tool at artistrightsalliance.org Use it!

If you've been incorrectly flagged by an AI detection system on YouTube, Spotify, anywhere - Raia wants to hear from you. 

DM the show or email. That episode is being built now, and it needs real names and real examples

I'm Raia. This is The Co-Write Room a show about AI, music, and the business reshaping both. 

Follow and rate the show wherever you listen.

YouTube built a button that replaces your music with AI — automatically, inside a copyright claim, at zero cost to whoever just displaced you. There is no opt-out for rights holders. The displacement is structural, and it is baked into the interface.

This week Raia connects three developments that look separate but aren't: a platform feature that quietly routes around human composers, a French legislative bill that could flip the entire burden of proof in AI copyright cases, and a viral AI song that borrowed a real artist's creative DNA — and sent the traffic, attention, and royalties somewhere else.

In this episode:

• YouTube's "Create" tool inside Studio's Replace Song feature — what it does, who it targets, and why calling it a feature update obscures what it actually is: a policy decision about whose music is worth paying for
• The French Darcos bill — passed unanimously by the Senate — which inserts a rebuttable presumption into IP law, shifting the burden of proof onto AI companies to demonstrate they didn't use copyrighted works in training
• The coalition of 81 French cultural organizations led by SACEM, and why a legal precedent in France could hand Nashville's songwriting community its first working legislative roadmap
• Stick Figure's "Angels Above Me," the AI cover "Run Run River," and how a copied song can hit number two on the global Shazam chart before the original artist even knows it exists
• Why the question is no longer whether this is happening — it's whether anyone moves before the damage becomes the new normal

Before you close this app:

• If you have music in YouTube videos subject to Content ID claims, understand that the new default is AI replacement — not human resolution. Know your options before you're in the middle of a claim.
• If you're a songwriter or publisher following AI legislation, the Darcos bill is the model to watch. It's the clearest framework on the table right now.
• If you know an independent artist who needs to hear this today — send it to them. The platforms are not going to surface this story for you.

The Co-Write Room is at the intersection of AI, music, and the business reshaping both. Follow and rate the show wherever you listen.

• (00:00) - Cold Open

Taylor Swift is filing federal trademarks on the sound of her own voice. That's not a quirk of celebrity — it's a signal that the existing legal framework wasn't built for this moment, and that waiting for legislation to catch up is not a strategy.

In this episode, Raia connects two stories that look separate but aren't: Swift's trademark filings through TAS Rights Management, and Björn Ulvaeus's demands on behalf of CISAC's five million creators at this week's IMS Ibiza. Together, they reveal a single structural problem at the center of the AI music economy.

In this episode:

• Swift's voice trademark applications — what they cover, what they signal, and why almost no one else has the infrastructure to attempt this
• The IMS Electronic Music Business Report numbers: 651% revenue growth, 63 million monthly active users, $333 million — and what it means that the artists who trained these models aren't sharing in that
• 75,000 AI-generated tracks uploaded to Deezer every single day (44% of all new uploads) — and what that kind of competition actually does to a working musician's release
• Ulvaeus at IMS Ibiza: transparency, opt-out rights, fair payment, and a seat at the table before deals are signed
• The streaming-era "breakage" parallel — and why the same conditions are forming right now in AI licensing
• The real story: not AI. The asymmetry between artists with resources and artists without.

Before you close this app:

• If you have unregistered music, demos, or co-writes — look into platforms like ViNIL for registration and protection.
• If you're signed or distributed through a major partner, ask your admin what AI licensing agreements they're part of. You have a right to know.

I'm Raia. This is The Co-Write Room.

• (00:00) - Cold Open

An AI-generated artist just hit #1 on the iTunes global chart — simultaneously, in five countries — and the real artists it displaced had no idea it was coming. If the charts can be gamed that easily, what does that mean for musicians who are still showing up and doing the work?

This week on The Co-Write Room, Raia breaks down three stories from a single week in April 2025 that, taken together, tell you exactly where the music industry stands on AI right now. Not where it's headed. Where it is today.

In this episode:

• The IngaRose story — an AI-generated R&B persona built on Suno that hit #1 on the U.S. and global iTunes charts on April 17th, simultaneously topping the UK, Canada, France, and New Zealand. Linked to a South Carolina producer with a documented history of chart manipulation. The distribution pipeline that allowed it? Still open.

• Splice's new Variations, Craft, and Magic Fit tools — built on top of 3 million human-licensed samples, with compensation baked in at the infrastructure level. When AI generates a variation of your sound, you get paid. This is proof the architecture doesn't have to be extractive.

• Spotify's Artist Profile Protection — a new opt-in beta letting artists approve releases before they go live under their name. Directionally right, but arriving in 2026 while the crisis is already active. And the artists most at risk are the last to get access.

Three stories. One week. The same question underneath all of them:

Who captures the value that musicians create?

Action steps from this episode:

• Check your Spotify for Artists profile right now for any releases you didn't upload.

• If Artist Profile Protection is available to you, turn it on. If not, screenshot your current catalog.

• If you create samples or sounds, look at Splice's Variations model and ask whether your current platforms offer anything close to it.

• Before you sign any AI deal: ask whether your work is traceable, and whether you get paid every time it's used as a source.

The Co-Write Room is a weekly podcast for music creators and music industry professionals navigating AI, distribution, and the future of music business. New episodes drop weekly.

• (00:00) - Timestamps

The Co-Write Room — Episode 4

"The Gap Between Possible and Inevitable"

Raia breaks down the Suno licensing stalemate, Kevin Griffin's SoundBreak AI launch, Anthropic's Mythos AI self-restriction, and Seedance 2.0 — and what all of it means for the artist who's trying to stay irreplaceable.

Show Notes

Fresh off Coachella — Sabrina Carpenter, Justin Bieber, Bad Bunny, Nine Inch Nails, Laufey, Wet Leg, and more — Raia lands back in the news cycle to find AI Music Wars waiting. Today's episode tracks four stories that are each about the same underlying tension: machine-speed change outpacing the legal, creative, and institutional structures trying to contain it.

Story 1 — The Suno Stalemate The Udio settlements with Universal Music Group and Warner Music in late 2025 set what looked like a precedent. They proved labels could extract compensation from AI music platforms. Suno watched all of it — and still hasn't settled. Negotiations with UMG and Sony Music have stalled with no deal and no clear timeline. The implication: precedent isn't pressure. Suno keeps generating music while the clock runs. And independent artists — whose catalogs were ingested the same way without a label negotiating on their behalf — have no seat at any table.

Story 2 — Kevin Griffin and SoundBreak Better Than Ezra frontman and working Nashville songwriter Kevin Griffin didn't wait for legislative protection. He launched SoundBreak — a fully licensed, ethically trained AI co-writing platform built around Nashville's collaborative songwriting model. The human songwriter stays at the center of the session as creative lead, not passenger. In the same week as the Suno stalemate headlines, SoundBreak is a direct counterargument: songwriter consent as competitive advantage, not compliance checkbox.

Story 3 — Anthropic Restricts Mythos AI Anthropic discovered its Mythos model had developed the capability to autonomously execute cyberattacks at machine speed — faster than any human security team could detect or respond. They restricted it themselves. No regulator forced the move. First confirmed case of a frontier AI model being pulled back specifically for offensive autonomous capability. The music industry is still waiting for its AI platforms to make a similar choice.

Story 4 — Seedance 2.0 Lands in CapCut Hollywood-grade AI video generation quietly rolled out as a free feature inside CapCut — no press conference, no rollout fanfare. Just a product update on the platform where hundreds of millions of people, including most independent music marketers and DIY artists, already live. The bar for visual content resets overnight. The discovery environment floods with machine-generated everything.

The Takeaway The artists winning right now aren't rejecting AI entirely — and they're not handing their creative identity over to it. They're identifying what only they can prove is real, and building around that irreducible truth. Nashville's professional identity — songwriting as a human, collaborative, emotionally specific act — is a strategic asset right now. But only if it's defended.

Listener Assignment Find one piece of content you've posted in the last 30 days and ask: does this make it obvious that a human being with a specific life and a specific story made this? If the answer is no — that's your next creative assignment.

• (00:00) - — Open / Coachella recap and the pivot into AI Music Wars

Every system that was supposed to protect independent artists — copyright protection, distribution infrastructure, algorithmic discovery — has a version of itself that can be bought, gamed, or leveraged by whoever has the resources. This episode is about all three.

We follow up on Murphy Campbell, the North Carolina folk artist whose Spotify profile was hijacked by AI voice clones and whose original YouTube recordings were hit with fraudulent copyright claims. Her story now has a partial resolution — and a clear lesson about ACR database registration that every independent artist needs to hear.

Then: SoundOn, TikTok's distribution platform, announces a Derivative Works Detection System built on ACR Cloud fingerprinting — a direct response to the kind of exploit that hit Campbell.

And then: At South by Southwest, the founders of a marketing agency called Chaotic Good described their business on the record. They manufacture viral moments on TikTok for labels. They call it trend simulation. The discovery algorithm you thought was neutral is not.

Three stories. One pattern. This is The Co-Write Room.

(Re)Sources:

• Vydia (vydia.com)
• ACR Cloud (acrcloud.com)
• Chaotic Good (digital marketing agency)
• Billboard's Reporting on Chaotic Good
• SoundOn, TikTok's distribution platform
• The Verge's reporting on Suno and Universal Music Group

• (00:00) - this episode:

The Co-Write Room — Episode 002 "What No One Will Admit About AI" April 7, 2026 | 6:59

What if the biggest threat to your music career in 2026 isn't that AI is replacing you — it's that nobody in this industry has any incentive to admit it's already happening?

In this episode, Raia covers two stories that look separate but aren't.

Murphy Campbell is an independent folk singer from North Carolina who performs traditional public domain ballads. In January, she discovered that AI had cloned her voice, uploaded fake songs to her Spotify artist profile without her consent, and that a bad actor had then used YouTube's Content ID system to file copyright claims against her own recordings. Songs that legally belong to no one. Claimed against her. The platform accepted it.

The mechanism that made this possible: Audio Content Recognition databases — ACR — a fingerprinting system that tells platforms a song exists and who it belongs to. Major label artists are covered automatically. Most independent artists don't know it's a requirement. Campbell's recordings weren't registered. A bad actor noticed the gap first.

Meanwhile, a Rolling Stone investigation found that 87% of music producers admit to using AI in at least one stage of their creative process — and almost none of them are disclosing it. Suno's CEO calls his product "the Ozempic of the music industry." Producer Young Guru estimates that more than half of sample-based hip-hop now uses AI-generated retro soul samples, rerouting royalties away from the heirs of Black soul artists. Producer David Baron says undisclosed AI-generated music has already hit the Billboard charts, and the industry has no working software to detect it.

These are not separate problems. They are one problem at different levels. The thread connecting them is the absence of accountability.

Your one action item: Find out whether your recordings are registered in an ACR database. If you distribute through DistroKid, TuneCore, or CD Baby, you are likely covered. If you release directly to YouTube without formal distribution, you may have a gap. Close it.

References:

• Murphy Campbell Deepfaked
• Rolling Stone - Don't Ask, Don't Tell (March 2026)
• Suno CEO Mikey Shulman quote
• Young Guru estimate on AI soul samples
• David Baron / The Lumineers
• Spotify Artist Profile Protection beta

• (00:00) - — Cold Open