Plagiarism in music is not a new phenomenon, on the contrary. Musicians and producers have numerous arguments for why they do this, from claiming that “everything is already invented” to “no one can have a patent on certain chords” and the like.
With the prevalence and availability of artificial intelligence, new negative trends in the music industry have emerged concerning counterfeiting, in the broadest sense. A song generated by AI that was released under the names Drake and The Weeknd is just one example.
However, artificial intelligence has also brought about positive changes within this creative industry, not only when it comes to the way it is expressed, but also the realization of the rights of music authors and interpreters.
How can AI detect musical plagiarism?
Those who regularly use Instagram or TikTok have probably already come across the profile of the creator of @itsjmaine, known for revealing plagiarism and original tracks in their videos. He confirms with its content that plagiarism is a very common phenomenon in the world music industry and that there is almost no known name, from Eminem to Metallica, which boasts a “clean slate”.
There are more and more users who similarly address this negative phenomenon, pointing fingers at musicians and holding them accountable.
For a more detailed explanation of how court disputes of this type are resolved, continue reading the linked blog.
In 2020, Katy Perry was accused of having her song “Dark Horse” plagiarized Marcus Grey’s “Joyful Noise” and had to pay a $2.8 million fine.
But in addition to the legal consequences that musical plagiarism entails, it seems that fans are much more affected by the emotional aspect that concerns the broken trust in the creativity of a particular author.
It could be said that a song that is found to be plagiarism of something existing automatically loses its meaning to some extent and implies the question "Is this original?" for each subsequent poem by the same author.
Identifying plagiarism in music using machine learning algorithms was dealt with by Rajesh Ramachandran Nair in a paper published in 2021, it points out that analyzing similarities in music is an extremely delicate and complicated procedure.
Online tools for detecting plagiarism in music
Today, there are a handful of online tools that can help musicians determine before releasing whether their song unintentionally sounds like something existing. These tools are also used by music enthusiasts to “prove their point.”
Some of the most common are:
- Grammarly: Yes, while this is a popular tool for checking for plagiarism in written content, it also offers opportunities to check for plagiarism in music. It can analyze the melody, harmony, and rhythm of a piece of music and compare it with other existing songs to find similarities.
- Audible Magic: Audible Magic is a professional tool for checking plagiarism in music intended for record labels and other business entities. It uses patented technology to identify similarities between songs and can be integrated into different platforms.
- AudioLock: AudioLock is a software for detecting plagiarism in music that works by analyzing the audio prints of a song. It can detect similarities and potential plagiarism even if the music is created with different instruments or in different styles.
- Plagiarism Checker X is a plagiarism detection software used to check for plagiarism in various types of content, including music. It uses advanced algorithms to scan and compare music files for similarities.
What do they have in common?
Almost all of them, more or less, use machine learning to analyze infinite databases and identify segments of songs that resemble each other. So, similar to seminar papers, these tools show how much of a song can be considered plagiarism.
In addition to the aforementioned tools, a good tip for musicians is to use song recognition apps such as Shazam, SoundHounds or MusixMatch to check if their song resembles an existing one.
How does Spotify use AI to detect plagiarism?
And finally, it should be mentioned that one of the largest streaming platforms long ago decided to put an end to musical plagiarism, devising its own technology for successfully and effectively detecting plagiarism.
In late 2020, Spotify released a patent titled “Plagiarism Risk Detector And Interface”, intended for music authors. It works by analyzing the sheet music scores that contain the basic elements of the song, including its melody, lyrics and harmony. The rhythm, as in most cases, is not analyzed for the possibility of plagiarism.
So, musicians can submit their own songs for plagiarism risk analysis, and Spotify detector then crosses the data with all the songs that exist in their database.
A new window will appear to the user that will mark parts of the song that appear in many songs, some songs, or are marked as brand new. In addition, all songs whose parts are recognized as identical will be linked so that the author of the track can be convinced of the similarities.
This detector greatly simplifies the process that would otherwise take much longer and require some manual checking. In addition, Spotify has one of the largest music recording studios, so authors can be sure of the reliability of the data, especially if they themselves planned to release music on the aforementioned platform.