We measure how music shapes AI — and make sure it pays.
AI music is built on human music.
Every AI-generated output carries the influence of the recordings and works it was trained on. Creative Weight Attribution™ makes this measurable, provable, and payable.
Our system analyses how individual tracks shape AI-generated output — capturing authentic creativity, relevance, and impact at the source.
Creative Weight Attribution™ accounts for both master and publishing rights — so every stakeholder in a recording is represented, from songwriter to performer.
Built to plug into the existing royalty infrastructure of collecting societies and rights organizations — run by a team that has handled royalty payouts for decades, globally.
Everything attribution has to deliver — measurement, consent, rights, insight and payout. In one solution.
A verifiable attribution record for every single training input and AI output — not just per model. Evidence-linked, audit-proof.
Influence measured where it happens: in training, at the prompt, and in the congruence of the output.
Opt-in and opt-out, enforced — with granular rules per work, per use and per context. Revocable at any time.
Master and publishing resolved per attribution — with consent structured to cover voice and likeness.
Know what your catalog is worth to AI — before, during and after licensing.
Dashboards and reports for rights holders and AI companies alike — from influence to payout.
Deterministic results with a full evidence trail — including EU AI Act transparency reporting.
Integrates anywhere. Embedded at a rights holder, at an AI company — or standalone in between.
What it does is on this page. How it does it — that's what a demo is for.
Book a demo →Documented in a published whitepaper, presented at DAGA, and developed alongside leading audio research. No black box — deterministic, reproducible, and open to independent verification.
Master and publishing rights from the ground up — because a fair AI economy pays everyone who creates a work: performers, songwriters, labels and publishers.
Designed to reward quality, uniqueness and creative impact — unbiased by catalog size or market power, so the artists who shape the sound of AI are the ones who earn from it.
Through our partnerships with rights management societies, attribution connects directly to the systems that already distribute royalties worldwide. Royalty distribution is what our team has done for decades, globally.
Creative Weight Attribution™ is developed and operated in Europe — designed in accordance with the EU AI Act, the GDPR and European copyright law. For rights holders and AI companies alike, that means a solution built for the world's most demanding regulatory environment.
Whether you're an artist, label or publisher: Creative Weight Attribution™ turns your catalog's influence on AI into a documented, recurring revenue stream.
Talk to us about your catalog →Fair, transparent, verifiable attribution is your license to operate. Show regulators, partners and the creative community that your model pays the music it learns from.
Book a demo →Where attribution creates value: licensing, rights resolution and payout — deployable at a rights holder, at an AI company, or in between.
What attribution runs on: audio, metadata, prompts, outputs and human resonance — identified, enriched and consented, from any source.
How influence is computed: deterministic, reproducible Creative Weights — from training, prompt and congruence.
Our team brings extended experience from across the music industry — independent labels, collecting societies, digital distribution and attribution technology. We have licensed rights, run royalty payouts and built the software behind them, globally.
Label operations, artist relations and rights management in the independent sector.
Rights administration and distribution with CMOs across continents.
Global delivery and royalty accounting for digital music at scale.
From audio research to production systems — the science and the software.
"Working with AIxchange on Creative Weight Attribution™ brings both worlds together: our comprehensive audio research and their hands-on music industry experience."
Fraunhofer is Europe's largest organization for applied research — the birthplace of the MP3. Its Institute for Digital Media Technology (IDMT), founded by MP3 co-inventor Karlheinz Brandenburg, is one of the world's leading institutes for audio analysis and music technology. Creative Weight Attribution™ is developed alongside this expertise — and presented to the scientific community at DAGA.
Our whitepaper, presented at DAGA, details how Creative Weight Attribution™ measures creative influence — and how that influence becomes revenue across master and publishing rights.
It measures the creative influence of individual works on what an AI model produces — combining three aspects: the training weight a work carries inside the model, prompt references to works, artists or styles, and congruences between the generated output and existing works.
Yes. Creative Weight Attribution™ measures creative influence on the merits of the work itself — quality, uniqueness and impact — independent of catalog size, label affiliation or market power. The model is deterministic and its results are reproducible and independently verifiable.
Yes. The model accounts for both master and publishing rights, so songwriters and publishers are included in AI revenue — not just recording owners.
Attribution results are designed to feed into existing royalty infrastructure. Through our partnerships with collecting societies, measured influence translates into standard distribution processes.
Creative Weight Attribution™ is developed and operated in Europe and designed in accordance with the EU AI Act (Regulation (EU) 2024/1689), the GDPR and European copyright law — including the transparency obligations AI providers face under the AI Act.
Yes. Creative Weight Attribution™ is built to work alongside licensed training data — talk to us about integration for your model and use case.
Creative Weight Attribution™ is ready for your catalog and your model. Whether you own rights or build AI — we'd love to show you how it works for you.
