AI and the future of the soundtrack: A panel with production:net and Hybrid Music Lab

AI and the future of the soundtrack: A panel with production:net and Hybrid Music Lab

 

AI and the future of the soundtrack

A panel with production:net and Hybrid Music Lab

From Mozart to Brian Eno, numbers and algorithms have inspired composers throughout history, as composers experimented with music that sounded as if they’d been created by machines. Today, the tables seem to be turning, as artificial intelligence allows machines to compose music that sounds convincingly human.

In the form of the soundtrack, AI-composed music is already creeping into the background of games, film and media. Will composers be outpaced by their artificial counterparts? Or does their human creativity make them irreplaceable? Will high-calibre composition become a site of human-machine creative partnership?

On Friday July 15th at Soho House, production:net and Hybrid Music Lab brought Berlin’s arts and technology community together to explore these questions with four expert panelists: Annette Gentz, CEO of Music & Film Arts management agency; Thomas Langhanki, game design professor and CEO of interactive storytelling studio Experimental Game; Valerio Velardo, CEO of AI composition tool Melodrive, and Fabian Mrongowius, producer at UFA LAB. Moderated by Creatives’ Catalysts founder AC Coppens, the discussion that emerged was challenging and, at times, heated — even as AI advances rapidly, the definition of creativity as uniquely human isn’t one we’re willing to give up so easily.

Definitions themselves are critical for this discussion, as the panel revealed from the outset. What do we mean when we talk about AI-created music? Composers in the audience were quick to point out that creating music has many stages, composition being just one of them. Bracketing out the role of new technologies and AI from music synthesis to performance (thought-provoking discussions in their own right), the panel zeroed in on the specific role of AI in music composition.

Much like debates around AI in other sectors, the first question raised is that of intelligent, automated technologies rendering humans obsolete. Looking at Valerio Velardo’s Melodrive, this doesn’t seem too far-fetched. Melodrive allows users to create intelligent and engaging soundtracks, even if they know next to nothing about composing music. What’s more is that Melodrive’s soundtracks respond intelligently to onscreen action, allowing audiences to become more deeply immersed. Having learnt from countless examples of different musical styles and their characteristics, Melodrive’s AI composes with sophistication. On one hand, Melodrive could be welcomed as democratization of sorts: with AI at their fingertips, anybody can build a moving and adaptive soundtrack, no matter their knowledge of music composition or sound design. 

On the other hand, the fact that an AI like Melodrive’s can do all the heavy-lifting threatens to put composers out of business. The existential ramifications of this were pointed out by artist manager Annette Gentz. Human-created compositions cost considerably more than those of a machine, making AI composers an attractive option for budget-minded film and game producers. And yet, many in the film, music and art worlds remain skeptical about whether AI-composed music can ever match the artistic quality and creativity of works composed by humans. Listeners too, it seemed, might be more likely to reject music as inauthentic if they knew it had been generated by a computer.

What would AI-composed music mean for the future of art on the whole? Members of the audience explored this question critically. The results of machine learning depend on what information we provide it with. If this input is more likely to be aligned to what people already know and love, we would run the risk of reproducing and cementing the status quo, leaving little room for unexpected and challenging artistic innovation. It seemed unlikely that AI composition could support creativity in the arts. Or could it?

Turning away from the notion of AI-as-replacement, the discussion considered how AI-composers might become tools or partners for human artists. After all, an AI benefits from the history of human musical creativity, not to mention the interdisciplinary collaborations between technologists and composers. Perhaps composers could benefit from having AI composition software in their creative arsenal. As composer and audience member Mona Mur ventured, AI could prove to be an exciting tool, with intelligent software offering new ways to unfold and build on human-composed motifs. In this way, AI might open a new chapter in the man-machine relationships shaping musical creativity, one part of a broader shift towards posthuman futures.