In my media class, our teacher asked us to write a song for our final project. For the first few songs students presented, every track sounded amazing. Then things got weird because everyone listening could tell that all the songs sounded similar. 

The teacher asked about our process for writing our songs. Only two people actually wrote the songs – everyone else had used AI tools.

AI generation in music might seem like a convenient shortcut to some people. However, as we learned from a discussion with our teacher, AI music is terrible. It is of poor quality, comes with a host of ethical and legal issues, and negatively impacts human artists, all of which we experienced in my class.

Quality

AI music is bad. First of all, AI-generated music relies on data collection, model training, seed input, and post-processing. Humans can customize the mood, genre, tempo, and instruments in the generator. While AI systems can change a lot of musical elements, like melody, harmony, rhythm, and timbre, they often rely on patterns from training data, which can make outputs sound similar and derivative.

The most important difference between humans and AI is that humans have feelings. AI music often feels robotic, predictable, and is less relatable because it lacks emotional depth and personal connection.

Sometimes the issues are technical. AI-generated music includes a lot of white noise in the track, like the hiss or crackle that people hear when listening to old radio or record tapes.

In other words, while AI music excels in efficiency and quick content creation, it fails to deliver an engaging listening experience or emotional resonance. This is reason enough to not listen to or use AI music, but if quality isn’t your concern, the next reason should be.

Ethical and legal issues

Laws regarding AI music are unclear. AI music is basically committing copyright infringement. Users don’t know which songs the generator used as samples because the original artists are not credited.

In addition to not crediting original artists, who owns the copyright to AI-generated music is still unclear. The users who input prompts into AI generators can’t claim the creation as their own, since they didn’t put much of their creative thought into it, but neither could those who created the AI tool claim copyright to the AI music.

Yotam Mann, co-founder of Never Before Heard Sounds, said “One risk of AI is that you end up with a song-generating machine that makes some beautiful compositions but it never credits any of the things it used to make.” 

Even if users use AI to generate a song, no matter how many ideas they put into it (lyrics, bpm, genre, etc.), they can’t release it without committing at least some copyright infringement, and sometimes enough to land in legal trouble. 

Impact on human artists

In my class, we debated whether the students who used AI generators should get the same grades as students who wrote the songs by themselves. 

When it comes to instrument arrangements, AI may outperform the students who wrote their own compositions. But those students invested months of work and don’t deserve to have their work so easily dismissed.

The same goes for professional artists. AI music reduces job opportunities for human artists, simply because it makes songs more quickly than people can.

However, excessive use and production of AI music leads to the homogenization of musical styles. While AI can mimic existing elements of music, it has significant limitations. AI doesn’t have the ability to creatively break the rules and bring true innovation. Hence, people should reject AI music to celebrate real artists who make creative and relatable songs. In other words, music should be for people, not for robots.

–

While AI music may offer speed and convenience, it comes at the cost of originality, emotional depth, and fairness to human artists. By relying on AI for music creation, we risk the diversity of musical expression and unresolved ethical and legal challenges. Therefore, we must prioritize human creativity and innovation over algorithmic convenience.