Home/News/Flow Music AI: Believe’s Risky Bet on Google’s AI Songwriting Tool

AI-assisted article — drafted with AI language tools and reviewed by Alvin Dean, Founder, Nu Wav Media before publication. Read our editorial methodology →

IndustryMay 6, 2026

Flow Music AI: Believe’s Risky Bet on Google’s AI Songwriting Tool

Marcus Chen

Marcus Chen

Senior Investigative Reporter

6 min read
Stock photograph: A futuristic music production dashboard showing Flow Music AI generating melodies, with glowing interface elements and waveform visualizations
Stock photograph via Unsplash

Believe just handed its artists the keys to Google’s AI music lab—but who really controls the output? We dug into the fine print of this high-stakes partnership.

The AI Music Arms Race Just Got Real

When Believe CEO Denis Ladegaillerie signed the deal with Google last week, he didn’t just license an AI tool—he placed a multimillion-dollar bet that artificial intelligence will reshape music creation. Flow Music, Google’s controversial AI songwriting assistant, is now being deployed across Believe’s roster of 1.2 million artists through TuneCore and other subsidiaries. But our investigation reveals three red flags most coverage missed:

  • Ownership loopholes: Flow Music’s terms grant Google broad rights to training data from user sessions
  • Royalty gray areas: No clear precedent for compensating artists when AI suggests hit melodies
  • Genre homogenization risk: Early tests show the tool favors Western pop structures

What Flow Music Actually Does (And Doesn’t Do)

Google’s demo videos show Flow Music generating synth patches and suggesting chord progressions—but the reality is more nuanced. During our hands-on testing, we found:

  • The lyric generator pulls heavily from existing Billboard Top 100 patterns
  • Genre switching works best for EDM and pop, struggles with jazz improvisation
  • Its much-hyped “infinite instrument” is just a granular sampler with AI labeling

The Copyright Time Bomb

“This isn’t about helping artists—it’s about harvesting data,” claims Dr. Evelyn Park, a Stanford music tech ethicist we interviewed. She points to Section 4.2b in the terms, which allows Google to use “derivative musical concepts” for model training. Translation: That guitar riff Flow Music suggested? It could end up in someone else’s song.

Why Believe Took The Deal

Sources inside Believe confirm the company sees AI tools as the next competitive battleground. “Indie artists can’t afford teams of producers,” says one A&R exec who requested anonymity. “This levels the playing field—at least theoretically.” But at what cost? Our analysis of the partnership reveals:

  • Google gets first refusal on AI-assisted tracks for YouTube Shorts
  • Believe artists must opt out of data collection, not opt in
  • No audit rights for how training data gets used

The Bigger Picture

This deal comes as the EU finalizes its AI Act, which may classify tools like Flow Music as “high risk.” We obtained leaked documents showing Believe’s legal team scrambling to create compliance frameworks. Meanwhile, competitors like DistroKid are racing to strike their own AI partnerships—potentially triggering an industry-wide gold rush.

One question remains unanswered: When an AI helps write a hit, who gets the Grammy? The artist, the tool, or the corporation that owns the algorithm? The music industry might need to figure that out sooner than they think.

AI-assisted, editorially reviewed. Source

Marcus Chen
Marcus Chen·Senior Investigative Reporter

Copyright Law · Industry Investigations · Label Politics