Provenance is becoming music’s new currency

Last week, Australia’s music industry did something no major market had done before: it changed the rules of its own charts. From the chart dated 31 August 2026, the Australian Recording Industry Association will no longer allow fully AI-generated recordings onto its official charts. Music that uses AI as a supporting tool — whether it be a drum pattern, a mixing assist etc., still qualifies. Music where AI generated the lead vocal or the core performance does not. To count, a track now has to be substantially human-made, and demonstrably so.

It’s a small rule with a big implication: provenance, defined as being able to show, not just assert, who or what actually made a piece of music, has become a formal, structural part of how the industry decides what counts.

ARIA’s move didn’t come out of nowhere. It followed a Brisbane DJ’s AI-assisted cover of a Madonna hit racking up tens of millions of streams and topping a chart before its AI vocals were even acknowledged. But it also sits at the end of a longer arc that runs back through 2025, when composer Max Richter gave evidence to a joint sitting of the UK’s Culture, Media and Sport and Science, Innovation and Technology select committees. Richter was careful to say he wasn’t anti-AI. His argument was narrower and, in hindsight, prescient: without a mechanism to attribute and reward the human work that trains and informs these systems, that work eventually stops getting made. He pushed for an opt-in model over the government’s proposed opt-out, and for real transparency about what AI systems had actually been trained on. The pressure from Richter and many others across the creative industries was part of what led the UK government to abandon its planned copyright exception for AI training this March.

Both stories are really the same story. Once you accept that AI can plausibly generate something “uncannily” close to a known artist, or top a chart under someone else’s name, the old assumption that origin is self-evident stops holding. You need a system that can actually demonstrate it.

The commercial data backs this up. IFPI’s Global Music Report 2026 recorded the industry’s eleventh consecutive year of growth, but flagged AI-driven streaming fraud as a direct threat to that revenue, and called for licensing models built around AI that properly compensate rights holders rather than working around them. In parallel, infrastructure for doing exactly that is maturing fast: the C2PA standard (Coalition for Content Provenance and Authenticity), originally built for images and journalism, is now being adapted for audio, embedding tamper-resistant metadata, including creator, source, licensing terms, directly into a file. Industry analysts covering this year’s Music Biz conference described data provenance as one of the core architectural patterns the music business is building under AI pressure, and predicted it as a template other content-dependent sectors will need to copy over the next eighteen months.

That’s the part worth sitting with. Provenance is being framed almost entirely as a defensive measure as proof against deepfakes, fraud, and the charts. But the same infrastructure that lets you prove a work is authentically yours is what lets you license it with confidence, price it accurately, and open new income streams from uses that would otherwise go undetected or uncompensated. Protection and monetisation turn out to be the same problem, looked at from two directions.

Music is simply further along than most other creative sectors in working this out. It has the chart infrastructure, the streaming data and the rights bodies to force the issue quickly. But there’s nothing music-specific about the underlying problem. Generative AI tools are now producing furniture concepts, packaging, typefaces and full product ranges with the same blurred authorship questions Richter raised about a melody: who made this, what was it trained on, and who’s entitled to be paid when it’s used commercially. Design is arguably next in line. Briefs are already being AI-assisted at the concept stage, and buyers, licensors and institutions will want the same kind of demonstrable answer the music industry is now building into its charts. Where music leads, design will likely follow.

This is exactly why we created CertifyIP, to help creative businesses in music, design and beyond build exactly this kind of concerted, evidence-based approach to their IP, treating provenance as something to build for, not just defend with. Whichever platform or standard your business ends up using, the direction of travel now looks fairly settled: if you can’t demonstrate where your work came from, you may increasingly struggle to get paid for it, or even to be counted.