Generative IA & farewell to creativity : are we so sure ?

I recently came across a drawing on social media that sparked my thoughts. It depicted a man lamenting that artificial intelligence (AI) had stolen his creative job, leaving him with more time for household chores—a shift he found unfortunate, as he would have preferred AI advancements to focus on these mundane tasks, thereby freeing him to pursue his creative profession.

However, I believe the opposite is happening thanks to artificial intelligence and generative AI: our creative potential is being unlocked. Let me explain: Whether it involves drafting a report, creating a painting, or composing music—a subject I am well acquainted with—the mastery of specific techniques and styles requires a significant learning effort, which can considerably hinder the expression of our creativity. As a composer and hobbyist, I spent years studying harmony, counterpoint, orchestration, composition techniques, and more. While these techniques are necessary to express creativity, they never helped me to generate musical ideas. They are akin to vocabulary and grammar, which allow us to translate and develop our ideas into a comprehensible language. Just as mastering the grammar and vocabulary of a language like English doesn’t write an original book, it’s the ideas and the compelling plot that make a work unique. Who hasn’t tried generating a report or an article with ChatGPT? Without an original idea in the prompt, ChatGPT merely churns out commonplaces with no originality. However, things get interesting when you precisely indicate all the ideas you want to develop and the editorial style—here, the story changes because its linguistic mastery takes over, resulting in a clear and finely crafted rendition of your ideas. This is a prime example of beautiful complementarity: humans having more time for the realm of creation, with the machine handling the application of techniques. This is true in painting as well: some criticize so-called artists who use AI to generate images, disparaging their work on the grounds that they used AI. But here too, it’s their creativity that speaks; it’s the originality of their ideas coded in the form of a "prompt" that holds value, with the machine taking over the technical execution. Creativity and techniques are complementary, and each has its value, providing an opportunity to express creativity for those who do not master the technique.

Similarly, in the case of a musical composition, an area I know well, I use a music notation software that includes features allowing me to save time on technical aspects so I can focus more on creation. I create chords and melodies and sometimes delegate to the software the orchestration of these chords based on given indications—techniques I know well but which are time-consuming and involve merely applying specific technical rules. I have also developed an iPhone app for harmonization that, based on a given melody, suggests a sequence of chords and optimal voice leading while respecting the hundreds of technical rules I know—a task that shows no intelligence but requires meticulous application of rules and is very time-consuming. Essentially, I've coded the automation of techniques to free up more time for creation. I think you get the idea.

Now, the question is about the future and sustainability of those who master these techniques. These are skills acquired through years of learning and often a vocation to turn into a profession. But what will their remuneration look like tomorrow? If techniques are particularly well-suited to learning by AI, what will be left for them? To imagine new techniques? As soon as these are learned, they are also acquired by machines... I don't have the answer, but we can reasonably be concerned about the gradual disappearance of these unique skills, often formalized in treatises but mostly acquired through the transmission of knowledge, analysis, and the study of those who had made it their profession.