The influence of artificial intelligence in music

Although we are often unaware of it, AI technology is already being used in many areas of our daily lives, including communication, the world of work, healthcare, transportation and entertainment. And it is very likely that this technology will increasingly contribute to solving various challenges facing our society in the future, such as climate change or increasing economic productivity.

However, everyone has been talking about AI since December 2022, when the US company OpenAI launched the online platform ChatGPT. ChatGPT understands what a user writes in text form, is capable of learning, answers questions or writes text passages with natural-sounding, almost error-free conversions in real time.

Rapid development

At present, however, development is still at an early stage, so the statements made by ChatGPT should be taken with extreme caution, as according to various AI experts, they are somewhere between "brilliant" and "breathtakingly stupid". However, given the rapid pace of development, it can be assumed that this will certainly change quickly.

And at the latest after it was announced in January 2023 that ChatGPT had passed a Master of Business Administration (MBA) at an elite university with a sufficient grade, there is a lively debate not only at universities and colleges about how to deal with this new challenge.

Solutions vary: where AI is not completely prohibited as an aid by the regulations, for example, its use must be documented transparently and the declaration of independence is supplemented by a note that AI was only used as an aid and not to find a solution.

Apart from the certainly exciting philosophical discourse about the origin of an idea (or in this case the solution to a problem) when using AI, proving that someone has not adhered to this requirement is likely to be quite difficult (at least at the moment).

For example, it is not yet certain whether advanced plagiarism software will be able to recognize whether a text was written by the author or generated by a "ghostwriter", which is what AI is in this case.

An examination board is most likely to find out whether a student's own work and thoughts are actually involved in a purely personal, scientific discussion (the so-called colloquium). 

AI in music studies

But how could AI - apart from its use in theoretical subjects of music studies or for writing a CD booklet or concert program - directly influence music?

In order to touch the audience with a live performance at a concert, I as a musician primarily need two things: emotions and a mature instrumental or vocal technique that allows me to express my feelings.

It seems likely that a machine will be able to copy, for example, an interpretation of a Chopin etude from a human model and play it to a high technical and musical standard.

But at best, this creates a good copy without an independent, artistic statement and vision. And let's not forget: the audience also wants to participate in the real lives of real musicians. 

What about composition? Today, there are already some composition machines (such as AIVA) that have been fed with countless music data and are able to analyze the key structures of music (i.e. rhythm, melody and harmony) and generate note sequences that are statistically the most probable - with sometimes astonishing results.

But statistical probability will only ever lead to mediocrity at best. AI will change many areas of music production and the fact that AI composes melodies that are used as background music or game soundtracks is already a reality today. 

AI and pedagogy

And will AI change pedagogical work? The fact that musical interpretations can be uploaded, analyzed by AI and then used to provide valuable suggestions seems to be a reality in the near future - not least because the research department at the Kalaidos University of Music is also working intensively on this development in cooperation with ETH Zurich, the pianist Ingolf Wunder and his company. Whether such a tool will completely replace personal music lessons seems very unlikely.

Successful pedagogy is about working with a student's individual abilities, weaknesses or feelings so that very personal musical statements and visions can emerge. Will AI ever be able to imitate these typically human characteristics?

Either way, even experts find it difficult to estimate what AI will be capable of. But music will probably never get by without humans.