Record Companies Want to Regulate AI Tracks. How Do Their Proposals for Charts and Streaming Services Compare?

The biggest AI music development of the summer is clear: The record industry feels it’s time to regulate AI songs.

In July, a coalition of music creators and the recorded music industry — including the RIAA, A2IM and SAG-AFTRA — came together to announce a proposal for how AI music could be treated across streaming services, aiming for a more uniform approach to regulation. Then, just weeks later, another record industry coalition, including many of the same players, suggested how charts should handle the inevitable rise of some AI-generated hits.

These announcements mark a significant shift for the record biz. Over the last few years, as AI music went from a rare novelty to an expected, if not always welcome, part of the online experience, the recorded music industry largely stayed out of the growing public discourse about what specifically should be done to mitigate this new reality. Should AI songs be tagged, removed or algorithmically penalized in some way on the web? Should they earn the same royalties and compete for the same charts as entirely human-made content? And there’s always the sticky subject of AI-assisted music, a wide-ranging category that could include anything from a human-made song with a three-second AI-powered guitar lick to a 99% AI-generated tune.

Even though the recorded music biz didn’t unite to prescribe any specific remedies for AI-generated music on charts and online platforms until now, the biggest labels’ general position on the effect AI music has had was already clear with the majors’ lawsuits against Suno and Udio in June 2024. In those, the trio claimed that AI music “saturate[s] the market” and “directly compete[s] with, cheapen[s] and ultimately drown[s] out the genuine sound recordings.” In effect, the labels argued, something must be done about the impact of AI music on the marketplace — but exactly what that something should be was unclear. 

Now, the record business is ready to speak up, and it seems the labels’ solution is just that — labels for AI music. The proposal hinges on stamping AI content on streaming services with a disclosure right beside it, much like a parental advisory sticker on a CD. The charts proposal, meanwhile, is more focused on curbing the appearance of AI-generated songs that were created with an “unauthorized” model, like Suno, which is trained on a wide swath of unlicensed music. For AI-assisted music to be chart-eligible, the proposal also would require it to be “substantially” human made and not raise stream or chart manipulation concerns. 

So how do these two proposals compare? 

They’re calling for different things

While both of these proposals deal with how to handle AI music, and while the solutions they argue for are broadly similar, they are asking for different things from different organizations. The trade groups are appealing directly to DSPs and asking for a labeling system, while the label groups are pushing for changes to music chart rules. In essence, one is about being clear and transparent for listeners as to what it is they’re listening to; the other is about what should be counted among the biggest songs and artists in the world on industry standard rankings. 

Both could be challenging to enforce 

The streaming service proposition suggests two labels: one for AI-generated music and one for AI-assisted music. AI-generated music is defined as songs with “lead vocal performance generated by AI, key instrumental performance generated by AI, or entirely prompt-generated AI music.” AI-assisted is defined as songs that use “generative AI… for some expressive elements. Humans performed the lead vocal and primary instruments.” 

The definitions for these categories are vague and open to interpretation, presenting a challenge to implementation. Plus, the proposition suggests that music uploaders should voluntarily disclose AI use, rather than the streaming services relying on an AI detector to flag the generated songs. But it seems likely that some users will lie about their use of AI, given that these labels seem more like a scarlet letter at this point than anything else.

The chart proposal has its own hurdle when it comes to enforcement. In the press release, the coalition notes that a song made with AI’s help must still be “substantially human made” to qualify for the charts. It’s unclear what meets that threshold of human creativity, and how the charts who implement these measures would check for it. 

The supporters are a little different

The record industry organizations that signed on to the AI chart principles and the AI streaming service principles are similar but still distinct. The key difference is that the streaming proposal was spearheaded by trade organizations representing record companies, and the charts proposal was directly pushed by a smaller — but powerful — coalition of individual record companies. 

Here’s who signed on to the streaming service proposal: RIAA, A2IM, Worldwide Independent Network (WIN), the European Independent Music Companies Association (IMPALA), the Recording Academy, SAG-AFTRA and the Human Artistry Campaign. This encompasses the vast majority of American and European music companies as well as some talent via the Human Artistry Campaign and SAG-AFTRA. 

Here’s who signed on to the charts proposal: Believe, BMG, Concord, Dirty Hit, Glassnote Records, HYBE, Mom+Pop Music, Partisan Records, Sony Music, Universal Music Group and Warner Music Group. (BMG, Concord, Dirty Hit, Glassnote and HYBE are not part of the organizations spearheading the streaming service proposal, although a few of the firms have distribution agreements with music companies that are part of them). 

The implementation

A key difference between the two proposals is that the chart principles are being implemented on some smaller global music charts, while the streaming principles are still in the proposal phase. Shortly after the chart guideline announcement, IFPI (an international trade organization for record labels) followed up with another press release, stating that the aforementioned chart proposal would be applied to a number of rankings worldwide, including the ARIA charts in Australia, the Official MENA Charts and the Official Southeast Asia Charts. Notably, it has not yet been formally adopted by the Billboard charts or the Official Charts in the U.K. (Many of the other “official charts” involved in the IFPI announcement are owned by IFPI, while the U.K.’s Official Chart is run by the Official Charts Company (OCC).)

While streaming services don’t seem to be opposed to the measures suggested by the record industry, none have yet adopted them. The Digital Media Association (DIMA), the U.S. trade body representing streaming services, has, however, signaled its openness to working with the labels on this proposal, saying: “DIMA has long advocated for the creators, owners and distributors of music to provide accurate and timely metadata on all music released and distributed to streaming services. We are following today’s announcement closely and look forward to receiving more detailed and accurate AI metadata, which will strengthen our ability to give fans the transparency they deserve. That information flows best when it travels the entire path from creator to fan, and our members rely on industry partners to make that possible. Our members look forward to continuing with the labels, producers, artists and distributors, as well as other industry stakeholders and standards bodies such as DDEX, to build a robust supply chain in which consumers can trust.”

Already, many streaming services have policies in place that are somewhat similar to what is being proposed by the labels, but there’s no unified rule book. Spotify and Apple Music have invested in building AI credits, meaning uploaders can disclose in the song’s metadata exactly how AI was used in the song creation process, but neither has gone so far as to have visible tags listed next to song titles. Deezer and Tidal both offer visible tags for AI-generated music, but they only do so for fully AI-generated recordings, not AI-assisted tracks, as the labels proposed. 

The scale of streaming services

There are a lot more songs on streaming services than there are songs on the charts, making the scale and difficulty of implementation vastly different between these two propositions. Rolling out a measure to curb AI music on charts with about 100 to 200 tracks could be relatively manageable, but streaming services are dealing with an influx of more than 100,000 tracks added daily, according to Luminate, and will need to set up complex infrastructure to make this work — if they even agree with the record industry’s proposal. Plus, given that AI songs are still largely unpopular, it’s unlikely, at least in the near term, that many of them will reach the major charts. 

While it’s impossible to truly know the exact number of AI-generated and AI-assisted songs added to streaming services daily, Apple Music chief Oliver Schusser told Billboard On the Record in April that fully AI tracks account for about a third of monthly uploads. Deezer also recently claimed that fully AI tracks topped 50% of daily uploads “at peak levels” in June 2026. 

To label or not to label?

The streaming service proposition largely hinges on the idea of adding visible labels next to AI-generated or AI-assisted tracks, but the idea of labeling is absent from the charts announcement. The press release for the latter does note that this “move follows the recent announcement by a broad spectrum of recorded music creators and companies around the world, of a unified approach to voluntary music track labeling to give fans clearer information about the use of generative AI (GenAI) in sound recordings,” but it doesn’t say anything about adding disclosures to the AI songs that do qualify for the charts.