Artificial intelligence (AI) and emotions
The use of artificial intelligence in music production has become a hot topic. On the surface, it’s about copyright law, but hidden within that is the accusation that it’s morally reprehensible for artists to make use of AI in production. Reason enough for a concerned person to take a stand on this. My name is Horst Grabosch and I am a book author and music producer at the Entprima Publishing label.
As a curious person, producer of electronic music and former professional musician and later information technologist, I have been involved with the use of machines/computers from the moment the technology had developed to the point where it was a useful aid. At the beginning it was basically about notation technology, then with the arrival of digital audio workstations about the production of demos and from 2020 onwards with the whole production chain of electronic pop music. So the use of machines is truly not a new field, and voices were heard early on condemning the use of electronics in music. Already earlier it was about the ‘soul of music’. Interestingly, these nostalgic critics hardly bothered with the analysis of what constitutes the ‘soul of music’ in the first place. The common listener didn’t care much, because he absorbed the feelings of the production as he personally found them in the production. A very wise decision, because in the chorus of the musical guardians of morals one found more and more absurd aspects, which called for damnation without any philosophical basis.
Since pop music is strongly influenced by stardom, listeners also sometimes missed a human idol behind the musical results,but this is merely a marketing aspect that has been completely compensated for by the arrival of DJ’s on the stages, at least in electronic dance music. As machine support became more widespread, thousands of amateur musicians saw their chance to produce music and publish it on the streaming portals. Of course, most of them couldn’t even fill a bathroom with fans, and so producers remained faceless. Faceless figures largely evade criticism, but some of them even managed to achieve tolerable success in the wholly new world of sound consumption driven by mood playlists. The many unsuccessful ‘learned’ musicians had envy written all over their faces. Many jumped on the bandwagon because, as trained musicians, it was of course even easier for them to produce electronically, but the sheer volume of productions meant that their works sank into no-man’s land. To make matters worse, artificial intelligence has now reached the point where it can produce complete songs, including lyrics on the fly. Desperation is spreading among producers who have not yet achieved sustained algorithmic attention, especially since it is to be feared that virtually anyone can throw songs onto the market. A vision of horror for all music producers.
Most listeners don’t even know what’s happening behind the scenes and they don’t really care, the main thing is that they continue to find enough songs for their needs, and there are now millions of them within their subscription models. However, these listeners are the target group of most desperate producers. They can now join the ever-growing number of mood sound painters, or produce songs with so much soul that they stand out from the crowd. They must stand out enough to compensate for both the lack of a real ‘face’ and the lack of a real character voice. The Japanese have already impressively shown how this is possible with artificial voices and avatars, which, however, required a lot of computing power and programming expertise and was accordingly costly. The rapid development of artificial intelligence has now opened up this construction kit, or Pandora’s box as some people think, for everyone.
It is up to us what we make of it. We don’t need to be afraid of AI, because it only does what producers have always done, emulate the successful models and possibly find new combinations in the process – only AI can do it in seconds. Producers who embark on this path must deliver extraordinary results, but didn’t they already have to do so in the “good old days” in order to be successful? So what’s so new in this respect?
It’s the path to the result, and therein lies the wonderful opportunity that AI-assisted music production brings us. As a producer, you no longer have to spend time learning genre-specific production details, because the AI can simply do that better, because it has analyzed millions of role models in terms of success. This means that you can focus entirely on your intention in terms of triggering feelings in the listener – and that has always been the intention of music. You have to shape and tell your story. Of course, that means you’re only partially putting the AI in the driver’s seat and never relinquishing responsibility over the outcome. Whether you then succeed with it depends on only two questions. Does the listener want to remain in the superficiality of habit, or is willing to engage with your story. In my opinion a very pithy and almost philosophical reduction of musical success factors. As far as advertising and marketing are concerned, almost nothing changes – almost. I jumped on the bandwagon of AI-assisted music with the advent of chatGPT, and may I point to the results, which have already been released as singles and will soon be released in full as an album. Myself, the songs have moved more than previously created. Given the intensity of my personal interventions in the songs, it wasn’t a time-saver (and thus authorship in terms of copyright is clear), but it has enormously expanded my toolbox as a storyteller and soul-searcher – and that’s why I’m sure to stick with it.