Hacked Suno source code reveals massive YouTube scraping, TIDAL becomes the first streamer to strip royalties from AI-generated music, and the record industry launches a voluntary AI labeling system.
Suno Hack Exposes Massive YouTube and Deezer Scraping
This is the biggest AI music story of the week. 404 Media published hacked source code from Suno revealing the company scraped 113,879 hours from YouTube Music (plus another 152,162 hours tagged separately), 62,117 hours from stock library Pond5, 12,287 hours from Deezer, and 17,615 hours from Genius — over 2 million music clips ingested in total. Suno used Bright Data's commercial proxy service with rotating IPs to bypass YouTube's anti-bot protections. The breach also exposed Stripe payment data for hundreds of thousands of users, with no breach notification issued.
Suno called the code "outdated" and "quickly contained," but the damage is done. This is exactly the kind of evidence UMG and Sony need for their ongoing copyright case in Boston. The RIAA's earlier allegations that Suno stream-ripped directly from YouTube now have source code to back them up. If you're building on Suno's API or using it in production workflows, the legal exposure just got harder to ignore.
TIDAL Strips Royalties from AI-Generated Tracks
TIDAL became the first major streaming platform to actively cut off royalties for fully AI-generated music, effective July 15. Flagged tracks get an "AI" badge, lose monetization eligibility, and are blocked from direct-to-fan sales. Music involving artist impersonation or fraud is banned outright. AI-assisted tracks — where a human led the creative process — keep full royalty access.
The move arrives alongside TIDAL's claim that streaming fraud costs artists roughly $2 billion per year. For producers using AI in their workflow, the distinction between "AI-generated" and "AI-assisted" is the one that matters. If you're writing the song and using AI for stems, mixing, or arrangement assistance, TIDAL says you're fine. If AI wrote the whole thing, you're out.
Record Industry Launches Voluntary AI Labeling System
Eight organizations — IFPI, RIAA, A2IM, WIN, IMPALA, the Grammys, SAG-AFTRA, and the Human Artistry Campaign — introduced a two-tier label for AI in sound recordings. A black square with white "AI" marks tracks where AI generated all or most creative elements. A white square with lowercase "ai" marks human-led tracks that used AI tools. The labels launched July 15 but remain voluntary with no enforcement mechanism.
The gaps are real: no streaming platform has committed to displaying the labels, and the definition of where "AI-assisted" becomes "AI-generated" is deliberately vague. The EU AI Act's Article 50 mandatory disclosure deadline on August 2 may force the issue across 27 member states, potentially superseding this voluntary approach entirely. Worth knowing about, but don't expect it to change your workflow yet.
Spotify Launches AI Chat for Music Discovery — Artists Push Back Immediately
Spotify rolled out "Talk to Spotify," a conversational AI assistant for Premium subscribers in beta across the US, Ireland, and Sweden. You can type or speak to it for music discovery, and it draws on your personal listening history. It maintains conversational context, so you can refine results ("make it more upbeat," "something like this but from the 90s"). Spotify says it uses a mix of proprietary AI and third-party models.
The rollout hit a snag within days. Lorde publicly called out Spotify's separate "About the Song" AI feature on Instagram after it generated inaccurate descriptions of her track "Current Affairs," attributing performance details from a completely different song. "I'm going to go out on a limb and say we don't want this," she wrote. Spotify pulled the text and said it "moves fast to fix" issues. The incident highlights the core risk of AI-generated metadata: one wrong detail erodes artist trust in the whole system.
Suno Brings AI Music Generation to iMessage
Suno released an iMessage keyboard extension that lets iPhone users generate 30-second AI music clips directly inside text conversations. You tap the plus button in a chat, select Suno, paste any message as a prompt, choose a genre, and send. Recipients need the Suno app installed to play clips. It's a novelty feature, but it makes Suno the first AI music generator embedded natively in a major messaging platform.
The strategic contrast with Udio couldn't be sharper. While Udio locks down its platform with DRM and blocks all exports, Suno is pushing into every surface it can reach — iMessage, developer APIs, consumer virality. Both approaches are bets on how the copyright cases will land. If you're evaluating which platform to build on, that philosophical split matters more than any feature comparison.
Warner Moves to Dismiss Session Musicians' AI Licensing Lawsuit
Warner Music filed to dismiss the AFM's lawsuit alleging that Warner and Universal licensed session musicians' recordings to Suno and Udio without compensating the performers. Warner's argument: it's not a signatory to the Sound Recording Labor Agreement, doesn't itself own the copyrights, and there's no existing AFM agreement covering AI training — so the payment clause "has nothing to point to."
This is the first time a major label has explicitly argued in court that no contractual framework exists to compensate session musicians for AI training use of their performances. If the court agrees, it would mean session players have no current legal mechanism to claim a share of AI licensing revenue — they'd have to negotiate new terms from scratch. The AFM's response filing will be the one to watch.
Spotify's new conversational AI assistant is the most significant AI audio tool launch this week — not because it's a creative tool, but because it changes how listeners find music. You can ask it questions in natural language ("play something like Boards of Canada but more percussive"), refine results through conversation, and query your own listening history ("when did I first listen to this artist?"). It's genuinely useful for music discovery in ways playlist algorithms aren't.
For producers and composers, the real implication is discoverability. If an AI chatbot is mediating between listeners and your catalog, the metadata, descriptions, and contextual information attached to your tracks matter more than ever. Understanding how these systems surface recommendations is no longer optional — it's part of the job. Available now in beta for English-speaking Premium subscribers in the US, Ireland, and Sweden.
Udio's BuyDRM deal this week crystallizes something that's been forming for months: AI music platforms are splitting into two fundamentally different models. Udio, post-settlement with UMG, operates as a sealed environment — you can create freely inside it, but nothing leaves. Multi-DRM protection (Widevine, FairPlay, PlayReady) plus a license-revocation tool means they can even claw back previously downloaded content. Meanwhile Suno is doing the exact opposite — pushing into iMessage, building developer APIs, chasing consumer virality and open distribution.
The split isn't philosophical, it's legal. Udio settled its copyright cases and accepted export restrictions as the price. Suno is still in active litigation and betting it can win — or at least build enough market share before the rulings come. The GEMA v. Suno verdict in Munich on July 31 could be the first domino. If you're integrating AI music generation into a production pipeline, the question isn't which tool sounds better — it's which model of access will still exist in a year.