Why is AI-Generated Music Controversial?

Image generated by DALL.E by OpenAI

What makes musicians so unique? Their ability to play an instrument? Their ability to read a music score? Or their ability to produce music through their imagination? If it is these qualities, then in a time when artificial intelligence can do the same, are they unique still? If musicians begin relying on AI to generate their music, a person with no musical background could be just as capable of making music.

By the late 20th century, the music industry began using AI to help musicians write and record more efficiently and creatively. The music scene changed again with the emergence of generative artificial intelligence, where creating whole songs from a simple sentence is now possible. However, unlike pre-existing AI, the response has not been as hospitable. The debate of whether this technology is doing more harm than good continues today. Yet, people are surprisingly unaware of how AI has changed the industry. Not knowing if the benefits outweigh the harm, musicians remain outraged about the introduction of generative AI in their field. While successful in many aspects, generative AI has not only threatened the safety of musicians’ jobs but has sparked a heated ethical and legal debate. A debate in which I find myself siding with the humans. So, why are AI music generators this controversial?

The achievements of artificial intelligence

It is no exaggeration to say that artificial intelligence is fully integrated into society, and the music industry is no exception. Within the music industry, technology has been integral in getting the sounds we know today. One of the earliest examples of such is autotune – a tool that allows artists to correct their pitch and change the sound of their voice. When Cher’s “Believe” became the first commercialised song to use autotune in 1989, the sound of pop music changed drastically. While not advertised, it is no secret that artists continue to use it. Many believe that it improves their compositions, making them more auditorily pleasing. When AI was introduced, the sound of music changed once again. The first forms of artificial intelligence were mainly machine-learning assistive programmes. Various AI helps streamline the mixing (the process of blending individual tracks) and mastering (the final part of audio production where the mix is polished) stage. A procedure that takes hours to perform can now be cut down to minutes, significantly improving efficiency and speeding up the release of music. There are even tools that can predict whether a song will be successful, allowing producers to tweak their compositions where necessary.

Recently, generative AI has made an emergence in the music industry. Generative AI is also a machine-learning model but with the ability to create something new. Within composition, programmes such as Amper and AIVA can make new melodies, chord progressions, harmonies, etc. Generative AI like Amadeus Code creates new lyrics and could be used in songwriting, especially useful for companies unable to procure teams of seasoned songwriters. With the existence of these models, the creative process of music production is more accessible and more easily completed, where even third graders can produce music. Nonetheless, the continued use of generative AI threatens the security of musicians’ jobs.

The takeover of artificial intelligence

The more AI being used in music production; the more jobs are at stake of being handed over to machines. The impact can be witnessed in another creative industry, when the Writers Guild of America protested against the unfair implementation of generative AI in screenwriting in 2023. The WGA eventually achieved a contract protecting screenwriters from being replaced by AI. Despite the success, warning bells have been ringing throughout the music industry.

Besides the songs we stream, music is present in the mundane parts of everyday life. It is in this part of the industry where people’s jobs and livelihoods are most threatened. For instance: in advertisement jingles, alarm and notification sounds, or elevator music. These all have humans behind them who have spent days composing and recording. When companies use generative AI, oodles of jobs will disappear, and breadwinners will find themselves trapped in poverty. Sure, new jobs may appear, but many believe that the number of new jobs will be minute compared to the number being fired.

Moreover, substituting humans with artificial intelligence eliminates the “human touch” – an element that is quintessential to anything creative. The finesse and flaws that only humans can provide will vanish, ultimately making music “too perfect” and “mechanical” sounding. In this sense, music would no longer be music.

What does this imply for current music producers? Since avoiding the loss of a plethora of jobs should be the goal, perhaps generative AI can be used as a co-creator rather than it being the sole creator. A suggestion: utilise AI to fill the gaps of human limitation. Technological progress is unavoidable; thus it is imperative that the industry grows along with it. Yet, there is still one more problem…

The ethical and legal predicament

One of the main controversies surrounding AI music generators is who actually owns AI-produced music?

When training AI music generators, a large set of mostly unlicensed music from the internet and music platforms is used. Allegedly, this violates the copyrights of millions of artists. Although, it is still legally ambiguous. It can be argued that, since AI generators are using small fragments of many existing pieces to create their own, owners of said pieces are being plagiarised and should be protected. On the other hand, many argue that the AI-generated piece should be considered “new”. Many also state that sampling other songs in one’s own song could be accused of the same crime. However, what makes sampling different is that it may only be done if explicit permission was given by the artist, writers and copyright owners. With this in mind, if creators of AI music generators get the permission of the millions of artists, writers and copyright owners of the millions of music used to train their AI, then they have a strong case against plagiarism accusations. Though, this is easier said than done. As such, the importance of transparency on music platforms has become emphasised.

Then comes the ethical dilemma. There are recorded instances where generative AI has been used in voice cloning and deepfake. One famous example is the faux Drake and The Weeknd song “Heart on my Sleeve” which went viral overnight. While it was eventually taken down, this incident proved the capabilities and advancements of generative AI; its ability to not only create an entirely new (and successful) song, but to also perfectly mimic the voices of professional singers. Due to this, many social media users were unaware that they were supporting fraudulent behaviour. This highlights the fact that, if used improperly, there could be serious ethical and legal ramifications. If not careful, we could arrive at an era where human expression and creativity is no longer praised, deepfake and fraud encouraged. This poses a significant question about the arts: does human creativity even matter when faced with technology that can easily and more efficiently do the same? Or does human artistry still overcome technology’s prowess?

As an avid music listener, I find the prospect of a machine taking over the music industry horrifying. An industry that thrives upon the input of human artistry, creativity, and expression without human input could hardly be called art. While the advantages of assistive and generative AI in music are clear, there is still much discussion regarding the potential social and economic harm. With the current situation, it feels like generative AI is more a threat than an asset. Companies wanting to replace musicians with AI and social media users using the artists’ voices for their own benefit screams capitalist greed and should be controlled. As human listeners, we should encourage organic creativity to support the qualities that distinguishes us from a non-living entity. Of course, at the rate technology progresses, avoiding the use of AI within the music industry could be impossible. But technological progress does not have to equate to human substitution. Considering how music has developed over time, it could be possible for AI music generators to be co-creators. Finding this balance will be a challenge, but it is a challenge and responsibility that we must share.