Alexis Lanternier’s Post

Sony, Universal, and Warner just drew a line in the sand on AI music. The three major labels, along with BMG and Concord, have proposed unified guidelines for how AI-generated tracks qualify for official charts worldwide. The requirement: music must be "substantially human-made," use properly authorized AI services, comply with copyright law, and show no signs of stream manipulation. This is not a ban. It is a framework for legitimacy. For AI music creators, this changes the equation. The guidelines create a clear pathway to chart eligibility for anyone using licensed AI tools and maintaining creative control. That means streaming revenue, chart recognition, and the ability to build a sustainable career as an AI-assisted artist. The industry is signaling that AI-made music can compete on the same stage as traditional production, provided it meets basic standards of transparency and human involvement. This is the legitimacy that creators have been waiting for. For studios, labels, and streaming platforms, these guidelines reduce legal and copyright risk while enabling them to support AI-assisted music without fear of fraud or unauthorized use. Chart integrity matters. Consumer trust matters. Platforms that adopt these standards early position themselves as responsible stewards of AI adoption, not reckless enablers. This is competitive advantage disguised as guardrails. The real signal here is that the industry is moving past the binary of "AI good" or "AI bad." It is moving toward a framework where AI is a legitimate creative tool when used ethically and transparently. The labels are not trying to kill AI music. They are trying to control it, monetize it, and protect their charts from gaming and fraud. But there is a tension worth noting. Several of these same labels have already signed partnerships with AI companies like Suno, Udio, and Nvidia. Some artists have reported facing non-negotiable AI usage clauses. The guidelines propose one thing. The business deals propose another. The question now is whether these standards will actually be enforced, or whether they will become a public relations layer over a more aggressive AI integration strategy. #AIEntertainment #HollywoodAI #FilmProduction #GenerativeAI #AICreators

The AI music problem isn't going away. It's accelerating. Deezer is now seeing 90,000 AI-generated track uploads per day. TIDAL just stopped paying royalties on wholly AI-generated music. Believe and TuneCore are blocking tracks made on Suno and Udio. Forbes estimated AI music fraud has become a $4 billion problem. I understand the impulse to ban. When I was running Spacial Audio and building the infrastructure for the early internet radio boom, we faced a similar moment. DMCA compliance, SoundExchange reporting — the entire webcasting industry was nearly shut down by blunt-force regulation before we built the systems that let it operate transparently at scale. The difference today is that the tools creating the problem are also being used by artists making genuinely original work. A creator who writes their own lyrics, crafts detailed prompts, uploads reference audio, iterates through multiple versions, and masters the final track — that's a different creative act than someone typing five words and hitting generate. But right now, streaming platforms can't tell the difference. The July 10 labeling program from IFPI, RIAA, A2IM, and others is a meaningful step — voluntary "AI-Generated" and "AI-Assisted" labels that bring transparency to sound recordings. But binary labels don't capture the full spectrum. There's a wide gap between a curator and a co-creator, and the infrastructure needs to reflect that. What I believe is needed: — Provenance that travels with the file — not a database lookup, but cryptographic metadata embedded in the audio itself — A deterministic scoring system that measures human participation — not a corporate black box, but an open-source standard the community governs — Transparency that works across all AI models, all platforms, all distributors — because this problem transcends any single company's ecosystem We're building toward this. Open-source. Community-governed. Model-agnostic. Not because we think we have all the answers — but because the alternative is a fractured landscape where every platform makes its own rules and artists get caught in the middle. I've been through this industry pattern before. The solution was never to ban the technology. It was to build the infrastructure that made it trustworthy. AI is the instrument. The question is whether we'll build the systems to prove who's actually playing it.

IFPI has rolled out new rules that will ban AI-generated music from official charts across its global network. Announced on 30 July, the principles now apply to charts in Latin America, the Middle East, Africa, and Southeast Asia, with 20-plus additional markets to follow. The proposal came from a coalition including Universal, Sony, Warner, HYBE, Believe, and BMG. The rules require three things: the AI service must be properly authorised, the track must be substantially human made, and it must not raise stream manipulation concerns. Note the word substantially. The labels have not defined what percentage of a track must be human-created to qualify. This is the sort of vague threshold that creates compliance risk for anyone sitting on the line. The commercial mechanism is about market access. Charts drive playlist placement, radio play, sync opportunities, and investor confidence. If a track is ineligible for charts, it is effectively invisible to the industry machinery that turns listeners into revenue. The labels are not banning AI music from streaming. They are banning it from the visibility ladder. For rights holders and technology founders, the practical action is to document your human creative input at the track level. Not the album level. Each recording needs its own provenance record showing which elements were human-created and which used AI tools, and which AI service was used. For investors and lenders, the signal is about who controls the definition. The labels and IFPI are writing the rules for what counts as legitimate music in commercial rankings. Catalogue valuations now need a line item for chart eligibility risk. puresignal.co.uk

Every major label just moved to keep AI music off the charts. Eleven companies — Universal Music Group, Sony Music Entertainment, Warner Music Group, BMG, Concord, HYBE LABEL, Believe, Glassnote, Dirty Hit, Partisan Records and Mom+Pop Music — proposed a set of principles governing whether AI-developed recordings should be chart-eligible. Eight days earlier, Deezer reported that fully AI-generated tracks had crossed 50% of everything uploaded to it on peak days in June. Roughly 90,000 songs a day. Here are the six criteria: • The generative AI service used must be authorised and lawful • The release must comply with copyright, related and personality rights • It must not breach the AI service's own terms of use • It must raise no stream or chart manipulation concerns • AI use must be signalled to consumers on downstream services • The recording must be "substantially human made" Five of those six test authorization, fraud, or disclosure. Exactly one asks how the record was actually made — and "substantially human made" is the only term in the document nobody defined. Now the number that complicates the panic narrative: those ~90,000 daily uploads convert to between 1% and 3% of Deezer's actual streams. Deezer says up to 85% of streams on fully AI-generated tracks were fraudulent in 2025. This is a supply-side flood meant to farm royalties. Ten weeks before signing the chart proposal, Universal licensed its catalog to Spotify's generative AI covers-and-remixes tool. Merlin — 30,000 independent labels, roughly 15% of the global recorded market — signed on Tuesday. On the same day's earnings call, Gustav Söderström said Spotify does not need deals with all three majors to launch, and that the next step is a public research preview used to gather preference data and train the model. Two documents, ten weeks apart, largely the same companies. One says AI recordings shouldn't chart unless the model was authorised. The other authorises a model. That shows this is the clearest strategic move the majors have made since streaming: they are saying "*we* authorize the models." A licensed AI derivative can chart. An identical-sounding unlicensed one can't. The line is drawn at the licence, not at the craft. Which means the fight over AI slop was never really about slop, but rather who holds the license when the tools get good. Worth being precise: these are proposed principles, not adopted rules. No chart body — Billboard, the Official Charts Company, anyone — has signed up yet. But three majors plus Merlin's membership plus a dozen prominent indies is most of the market. Chart compilers rarely say no to that. The charts aren't closing to AI music, they're closing to AI music the labels haven't licensed. As I've said in previous posts, licensed doesn't always mean ethical. Let's keep our eyes on this.

Introducing $ound/proof. Music has a credit problem. Every track is a web of contributions: a topline here, a beat switch there, a sample flipped three times over— and almost none of it is captured in a way that survives past the session. As AI accelerates how music gets made, that gap is about to get a lot wider. We're building the attribution and provenance layer for music: verifiable proof of who made what, starting from day one of creation. This is the beginning. If you're an artist, producer, engineer, or you just believe creators deserve credit for their work— we'd love to have you along for the build. Most tools try to detect what's AI, scanning a finished track and guessing. We prove authorship instead: every contributor discloses and verifies their own part, human or AI-assisted, and it seals into a tamper-proof credit certificate the moment everyone confirms. Detection classifies.  $ound/proof proves. For artists & producers: Get credited and paid for exactly what you did. Your contribution, self-attested and sealed from the first session, and yours to carry anywhere. For labels & rights holders: Verifiable, admissible provenance for every track: who did what, what role AI played, disclosed at the source— not guessed after the fact. For platforms & API partners: A provenance layer you can build on. Attribution and AI-disclosure data, verified at creation, ready to integrate. We are building toward a future where every creator is credited and paid for their work. Say hi to $ound/proof: the patent-pending attribution and provenance layer for music. Learn more at https://soundprf.tech/ (full demo best viewed on iPad/tablet or laptop/desktop, mobile & app soon!) Patent pending, U.S. App. No. 19/683,746. Follow $ound/proof here on LI! THEN Find us on Instagram (the page is brand new. FOLLOW US! We deeply appreciate your support): @ihavesoundproof In regards to our stance on AI & Music - Create w/Integrity: Make freely. Credit Properly. Disclose Honestly. Seal Permanently. Done right, we protect the one thing we have that connects us regardless of our differences. (see our latest post) #musicrights #independentartist #detroitmusic #musictech #creatoreconomy #attribution #provenance #detroitmusic #aiandmusic #aientertainment #aidisclosure #certificates #proof #credits #musiccredits #collaborators

More than half of the new music uploaded to some streaming platforms this year wasn't made by a person. That's not a headline meant to scare anyone off AI. It's the opening of a piece I just wrote about a boundary I think the music industry, streaming platforms especially, needs to draw for more clearly: not AI vs. human music, but AI music and human music, in separate rooms. There's a term used in the AI art world for this: deslopification. It's not about banning a technology. It's about not letting AI-made and human-made work quietly share the same space, the same chart, the same recommendation slot, the same royalty pool, without anyone being told which is which. For Harmonize, this connects directly to why we exist: music discovery built around real people, not algorithms guessing what you might like next. As platforms like Tidal, Apple Music, and Deezer experiment with AI labels and royalty rules, I think it's worth asking what boundary actually serves listeners and artists, not just what's technically possible. Full piece linked below. Curious where others in music and tech land on this. #MusicIndustry #ArtificialIntelligence #MusicTech #FutureOfMusic #MusicDiscovery #AIMusic

Google DeepMind launched Lyria 3.5 today, its newest AI music generation model, rolling out inside Google Flow Music. The upgrade focuses on more expressive, emotionally nuanced vocals with improved pronunciation, richer and more complex melodic arrangements, and tighter creative control over tempo and track length. It follows Lyria 3's February debut and Lyria 3 Pro's March expansion to full three-minute tracks. The more interesting detail for filmmakers is what Flow Music does beyond song generation. The platform integrates directly with Google's Veo video model, letting users direct an AI-generated music video for a song from the same platform, controlling characters, aesthetics, and visual details without a production crew. That positions Flow Music as one piece of Google's broader creative ecosystem, alongside Imagen for stills and Veo for video, rather than a standalone music tool. Flow Music is free to start with daily generation credits, with paid tiers mapping to existing Gemini subscription levels. https://lnkd.in/es2Z6T2r

【As AI lowers the barrier to turning music into a finished form, the real value shifts to experience】 According to Deezer, at its peak in June 2026, more than 50% of the new tracks delivered to the platform were fully AI-generated, reaching approximately 90,000 tracks per day. With generative AI, almost anyone can now turn an idea into something that sounds like a finished piece of music without spending a great deal of time. It is becoming increasingly easy to create a draft, refine it, and release it within a short period. This is a major expansion of what is possible in music creation. At the same time, as it becomes easier to give music a finished form, the music itself becomes harder to differentiate on its own. That is why I believe experience will become even more important. Game music is a clear example. The music that plays during a boss battle. The track that accompanies a major turning point in the story. The song you hear while attempting the same stage again and again. Game music is not simply background music. Its value comes from being connected to characters, visuals, interaction, narrative, and the player’s own memories. Even the same piece of music can carry a completely different meaning depending on where it was heard and what the player experienced at that moment. In the AI era, it may not be music itself that loses value. It may be music without context. The key question will no longer be how many tracks can be created, but how music can be delivered as part of a meaningful experience. How can it connect with games, characters, live events, visual content, and communities? How can it become something people remember? As the barrier to giving music a finished form continues to fall, the value of creative direction, world-building, production, and IP design will continue to rise. Reference: Deezer Newsroom https://lnkd.in/gEy9uA8s #GenerativeAI #GameMusic #MusicTech #MusicBusiness #IPBusiness

Suno just put a price tag on legitimacy. Starting September 3, the AI music platform is capping downloads: 7 for free users, 20 per month for Pro subscribers, 60 for Premier members. This is not a technical limitation. It is a business model reset. The move stems from Suno's settlement with Warner Music Group and reflects what streaming platforms have been screaming about for months: AI-generated music is flooding their systems, often improperly attributed to human artists or designed to game their algorithms. Spotify, Apple Music, and others have been forced to build spam filters just to keep their catalogs clean. Suno's download caps are the company's answer to that pressure. Watermarking is coming too, to identify AI-generated tracks at the source. This matters because the industry is moving past the binary of "AI good" or "AI bad." It is moving toward attribution, compliance, and revenue sharing. The real signal here is that AI music tools are no longer positioning themselves as free-for-all platforms. They are positioning themselves as legitimate production infrastructure. For AI creators, this is a hard pivot. The free tier is now a sandbox, not a distribution channel. If you want to generate music at production velocity, you need to pay. But here is what changes: the watermarking and label-backed models signal that the market is moving toward legitimacy and licensing. Creators willing to work within compliance frameworks now have a clearer path to revenue. The question is whether the tier pricing makes that path accessible. For studios, labels, and production companies, this is a competitive advantage. Suno's label-backed models promise faster generation, higher fidelity audio, and more control. The download caps reduce the risk of AI-generated spam polluting your catalogs or your artists' pages. The watermarking creates a clean signal in the marketplace. This is infrastructure that works with the industry, not against it. The larger shift: AI music tools are no longer trying to disrupt the music industry. They are trying to integrate into it. That requires compliance, attribution, and a business model that shares revenue with the gatekeepers. Suno is betting that creators and studios will pay for legitimacy. The next question is whether the pricing is right. #AIEntertainment #HollywoodAI #FilmProduction #GenerativeAI #AICreators

English Part one The AI Music Era: Why Suno and Udio Are Liberating the Industry Major record labels (Universal Music Group, Sony Music, Warner Music Group) claim that generative AI music tools destroy human creativity. History proves the exact opposite. Suno and Udio are the necessary and inevitable response to an industry that suffocates innovation through centralized distribution, homogenous algorithms, and insurmountable legal barriers. 1. The Centralized History of Major Labels Historically, the music industry has functioned as a rigid oligopoly backed by a hermetically sealed distribution infrastructure. The Gatekeeper Phase: Industry executives (such as Clive Davis) held an exclusive monopoly on who could access recording studios and physical store shelves. The Algorithmic Trap: With the rise of streaming platforms, power shifted to centralized, curated playlists and recommendation algorithms. The Homogenization of Sound: Major labels now optimize music strictly for the first 30 seconds of a stream to maximize payout metrics. This results in risk-averse, paint-by-numbers pop music that actively filters out creative anomalies. 2. Why Accessibility Accelerates Creative Quality The elitist argument that lowering the barrier to entry dilutes quality is historically incorrect. Every wave of technological democratization has expanded and elevated global creative output. [The Evolution of Accessibility] Physical Studios (Expensive/Closed) -> Home Recording/DAWs (Middle Class) -> AI Tools (Universal/Open) The Bedroom Pop Revolution: In the 1990s and 2000s, Digital Audio Workstations (DAWs) replaced million-dollar analog studios. This directly birthed modern electronic music, hip-hop production, and bedroom pop. The AI Democratization: Suno and Udio remove the remaining physical and financial barriers—such as the need to master complex software or purchase expensive instruments. A Shift to Conceptual Brilliance: The bottleneck shifts from technical execution to conceptual creativity. The human mind now operates as the ultimate curator, director, and visionary of the sound. 3. Why Unique Talent Always Rises to the Top Even when the entire world has access to the exact same AI models, true artistic talent will always stand out. The tool generates the audio, but the human injects the soul. The Prompt as an Instrument: An untalented user will merely generate generic pop clichés. A true visionary uses advanced prompting to fuse contrasting genres, weave obscure poetic lyricism, and enforce unexpected rhythmic structures. Iterative Curation: The creative process with AI relies on hundreds of micro-decisions. Talent manifests in the ability to select, modify, chop, and refine that specific 1% of generations that emotionally resonates. The Monopoly on Taste: Technical execution can be automated; unique human taste, lived experience, and cultural intuition cannot.