Here at Expansion Records, we have watched the rise of generative AI in music with a mixture of fascination, frustration and, at times, genuine alarm. We have seen the conversation swirl for years: around training data scraped without consent, AI-generated tracks flooding streaming platforms, deepfakes of established artists, and the uncomfortable question of what “authenticity” means when software can produce a hit-sounding record in seconds.

So when Music Week reported this week that major labels and independents have come together to propose a unified framework for AI chart eligibility, our reaction was immediate and unambiguous: it is about time.

Sony Music, Universal Music Group and Warner Music Group, joined by Believe, BMG, Concord, Dirty Hit, Glassnote, Hybe Corp, Mom+Pop Music and Partisan Records, have proposed principles to govern whether recordings developed using generative AI can qualify for inclusion in official music charts.

If implemented, this would establish an important boundary between human-led creativity and purely synthetic content. We welcome it-loudly and unreservedly. But welcoming the framework does not mean pretending the road ahead is straight or smooth. The challenges embedded in this announcement are significant, and they deserve to be named honestly.

Defining “substantially human-made” is harder than it sounds

The most consequential phrase in the entire framework is arguably the simplest: “substantially human-made.” Three words. Enormous implications.

What percentage of human creative input qualifies as substantial? If a producer uses a generative AI tool to create a full instrumental bed, then writes lyrics and records a vocal over the top, is that human enough? What if the vocal melody itself was AI-suggested? What if a songwriter uses AI to draft three verses and edits one of them into a chorus?

These are not hypothetical edge cases. They are increasingly part of the everyday reality of how emerging artists and producers work. Music creation has always been collaborative, and the boundary between “tool” and “creator” has always been contested-from drum machines in the 1980s to Auto-Tune in the 2000s.

Generative AI intensifies that ambiguity to a degree that a three-point framework cannot fully resolve on its own. The industry will need robust, transparent criteria that can be applied consistently. Those criteria will be tested-and contested-once they are applied to real releases.

Policing “authorised” AI models opens its own can of worms

The requirement that only properly authorised AI services qualify tracks for chart consideration is sensible in principle. Litigation is ongoing against multiple AI companies accused of training models on copyrighted music without consent, and charts should not legitimise work derived from stolen creative labour.

Enforcement, however, raises difficult practical questions. How will chart compilers verify which AI tool was used during production? Will artists and labels be required to disclose this when a track is submitted? If disclosure is voluntary, what prevents bad actors from simply omitting it?

The industry’s parallel initiative on voluntary track labelling, announced just weeks ago, is a step in the right direction. Yet the word “voluntary” carries its own warning. Until disclosure is mandatory and independently verifiable, the authorisation requirement risks being honoured more in the letter than in the spirit.

The major labels and independents backing this framework will need to push chart compilers and streaming platforms to implement verification mechanisms with meaningful consequences.

Streaming manipulation remains an industrial-scale problem

The third pillar of the framework-that qualifying tracks must not raise streaming or chart-manipulation concerns-points to one of the most immediate crises facing the industry.

Deezer alone has reported logging 75,000 AI-based tracks per day on its platform. Much of this content is suspected of being deployed specifically to exploit streaming royalty systems, with fraudulent stream counts inflating payouts and distorting chart positions.

This is not a fringe problem. It is a structural threat to the integrity of charts that artists, managers and fans still largely trust as a reflection of genuine public engagement.

If adopted by chart compilers worldwide, the proposed principles would help exclude the most egregious offenders from official recognition. But chart-eligibility rules cannot eliminate streaming fraud by themselves. That requires action at platform level, improved detection technology and, ultimately, stronger regulatory oversight.

Unity among companies does not yet mean unity across the industry

There is a notable tension worth acknowledging. The same Music Week report that broke this story also referenced a separate and less harmonious moment from last month, when artists, songwriters and managers publicly expressed concerns about the misuse of their rights in AI licensing deals being struck by the very companies now proposing these chart principles.

That disconnect matters. A framework designed to protect human creativity carries less moral authority when creators feel sidelined from the deals governing how their work is used to train AI in the first place.

For this initiative to have lasting credibility, it cannot remain a label-led effort. Artists, songwriters, producers and their representatives need genuine input into how these principles are defined, refined and enforced.

Protecting chart integrity is important. But it is upstream of a deeper question: who owns and controls the creative legacy that makes the charts worth caring about?

The bottom line

The announcement is genuinely significant. For the first time, the music industry’s major and independent players have found common ground on where the line sits between human artistry and synthetic content in the context of chart recognition.

That consensus has been a long time coming, and it sends a meaningful signal to chart compilers, streaming platforms and AI developers alike.

But drawing a line is not the same as holding it. The challenges ahead-from definitional ambiguity and enforcement gaps to unresolved tensions with creators-are real and will require sustained, collaborative work.

At Expansion Records, we will be watching closely, advocating loudly for artists and holding everyone accountable for what comes next.