The issue with the “AI versus human” music narrative

Whereas attending the NAMM convention and Grammys week occasions during the last fortnight, AI was, after all, a dominant matter. AI-assisted devices shared house with conventional bodily devices at NAMM; on the Grammys, Recording Academy CEO Harvey Mason reaffirmed that “solely human creators” are eligible for awards, whereas award winner Jon Batiste urged viewers to “defend our humanity in music proper now”. 

In these moments I’ve typically questioned: A decade from now, what’s going to we take into consideration the conversations round AI that we’re having right this moment?

Undoubtedly, there will probably be components of AI’s disruption that we couldn’t have seen coming. However one future reflection strikes me as extra apparent. The panorama of AI for music creation is extremely nuanced – but by comparability, the best way we speak about it will probably really feel black-and-white.

I generally name this the “AI unhealthy, human good” strategy.

For instance, conversations about policing “AI-generated music” on streaming platforms typically low cost the truth that the traces between AI-generated, AI-assisted, and “human-created” music is blurring. Audio system at NAMM had been fast to level out that solely “human-created” music is an oxymoron, as AI has been embedded in music creation software program for years. In opposition to this backdrop, initiatives like Bandcamp’s latest AI-generated music ban are well-meaning, however lack nuance and implementation technique in methods that can in the end trigger extra points than they clear up. Already there are remark sections stuffed with disgruntled, real-life musicians whose work has been unfairly faraway from the platform by customers who flagged it as suspicious.

Maybe much more harmful, although, is the danger of failing to recognise the numerous real-life, human artists who’re utilizing AI to invent completely new types of inventive expression.

Take Portrait XO, an audiovisual artist and researcher whose WIRE album and stay efficiency treats AI as a collaborator. Or Arca, whose artwork explores the fluidity of self, and who has used machine studying to warp vocals, construct efficiency avatars, and create music that mutates alongside its surroundings for the Museum of Fashionable Artwork. Or Holly Herndon, one other artist and researcher who explored the idea of id by constructing her personal vocal instrument, Holly+. As Herndon places it in a New Yorker profile:

“I feel it’s smart to be cautious… I feel [AI is] going to unleash an countless hose of s***** media. That’s 100 per cent going to occur. I simply don’t assume that’s the one factor that’s going to occur.”

These artists are inventing new types of expression – and certain view their work as each bit as inventive as they might a musician toiling away on a bodily guitar. In fact, these AI-forward artists are at one excessive, however there’s additionally an enormous center floor of artists, producers, and songwriters for whom AI is already now embedded as an assistive software. Finally, to deal with all music involving AI as damaging and unworthy of listening is to do them a direct disservice. In a means, the identical initiatives which can be supposed to guard human artistry are generally neglecting the identical folks they purpose to help. 

(Curiously, this raises new authorized questions, too. Within the US, AI-generated work can’t maintain copyright. But because of this if an artist creates their very own mannequin solely educated on their very own work – as Portrait XO did for WIRE – any output doesn’t belong to that artist. If a core tenet of copyright legislation is to incentivise, and thus foster, creativity, this seems like a blind spot.)

As soon as once more, a favorite 2024 essay by the author Katherine Dee places issues into perspective. Dee argues that critics typically fail to anticipate the long run as a result of they mistakenly search for indicators of right this moment’s tradition in the identical locations the place yesterday’s tradition was discovered. As a substitute, she means that there are new types of cultural expression and storytelling rising that we don’t totally perceive but. On the time, Dee used TikTok sketch comedies for instance. This format, which the leisure business didn’t take critically in 2024, has not too long ago been recognised and given a flashy new time period: micro-dramas.

Artists’ and business executives’ considerations concerning AI are completely legitimate, and this disruption ought to immediate a reassessment of how music is categorised, surfaced, and monetised. But it’s the underlying infrastructure that now not works, not the artwork kind itself. It’s important that we depart room for the gray zone – the distinct risk that on the crux of AI and music are completely new inventive expressions. We simply won’t have the identify for, or understanding of, them but.

The excellent news is that this realisation solely additional bolsters the necessity to invent new licensing options for AI  – an crucial which the music business is, by now, largely in settlement on. But there’s a twist. This isn’t nearly begrudgingly coming to the desk, to keep away from being on the menu. Additionally it is about taking an offensive, proactive lead: Empowering artists to invent completely new types of creativity, ones we don’t have the phrases for but, and to forge forward figuring out they will accomplish that in a means that’s accountable and equitable to the supply materials. As a substitute of enjoying fixed protection, it’s about recognising that human artists and AI are sometimes on the identical group