Blog Music

UMG, Endel, and the case for labels in the AI space

Cover image for UMG, Endel, and the case for labels in the AI space

Photo: Image generated using DALL-E

Photo of Tatiana Cirisano
by Tatiana Cirisano

The flood of new music created using artificial intelligence (AI) will further erode major label market share, and some label executives’ recent comments about the value of human-made music seem to put those fears in plain sight. No doubt, AI will massively drive up the amount of music that is made and released. However, there is no reason why labels cannot share in this new frontier. In fact, there is all the reason that they can, and probably will.

Universal Music Group’s (UMG) partnership with generative music app Endel, announced last week, helps paint this picture. The deal enables artists on UMG’s roster to create “functional music” soundscapes, including both new releases and new versions of old albums, which will live on the Endel app and be released as static albums on major DSPs. Whereas Warner Music Group CEO Robert Kyncl recently decried the idea that “an Ed Sheeran stream is worth exactly the same as a stream of rain”, Endel would seek to create an Ed Sheeran rain stream. This is likely just the beginning of a new wave of major label deals in the AI space.

Setting the record straight on consumer attitudes

First, let us debunk the narrative that consumers feel overwhelmed by the amount of music on streaming services, and the related idea that they simply may not want to listen to AI-assisted or AI-generated music. According to MIDiA’s Q4 2022 survey, 36% of music streamers agree that they struggle to navigate the vast amount of music on streaming services. That is not nothing, but it is far from the majority, and close to a quarter of music streamers are neutral on the matter. All of this means that the majority of consumers seem to not mind oversaturation. Further, 71% of music streaming weekly active users say they like being offered AI-generated / illustrated content on platforms like Spotify and Audible, according to MIDiA’s Q1 2023 survey.

There is little point in trying to change consumer behaviour. It makes sense, then, for labels to take the “if you can’t beat ‘em, join ‘em” approach.

Labels can own the AI space — but have a responsibility to creators

Major labels own much of the iconic music catalogue that AI developers would probably like to train their tools on. AI will surely increase the volume of music out there, eroding at labels’ market share, but labels can vacuum some of that share back up by guarding their datasets and licensing them out strategically. With its Endel deal, UMG is demonstrating that it is open to working with AI music companies that comply with licensing — and soon, we can probably expect UMG to make an example of a company that is not. Labels will have to adapt quickly to new technology to own the AI space, and while this has not typically been their strength, most now seem keen to change that.

However, the onus is on labels to work with platforms that can ensure artists and songwriters are fairly compensated, and that revenues do not flow disproportionately to the labels themselves. This is not a path to rake in free market share that labels do not have to pay creators for. If handled well, AI could open up a range of new revenue streams for artists — and ones that actually stem directly from their music (imagine that)! While all music releases naturally have a diminishing half-life, licensing that music to datasets is a way to extend that half-life.

On that note, creators may have new leverage in the AI space. That is because the “rights of publicity” that would likely be implicated in some AI music cases — such as an artist like Grimes licensing their vocal for use in others’ music — belong to artists, not labels. For the same reason, labels may not be able to initiate lawsuits based on the misappropriation of their artists’ voices — artists would need to do so themselves. Label contracts will likely evolve to include clauses related to artist deepfakes and other AI uses, and artists should hold tight onto their leverage in these conversations.

If you enjoyed the themes in this blog post, check out MIDiA’s new report, “AI and the future of music: The future is already here”, and be on the lookout for additional reports on AI in entertainment over the coming months.

The discussion around this post has not yet got started, be the first to add an opinion.


Add your comment