For my final project, I explored AI ethics and music, specifically as it pertained to copyright infringement, right of publicity, and fair use. Here I included a sample of an AI generated song by Musicfy. I used Miley Cyrus’ “Party in the USA” as the song. In some parts of the audio, I uploaded clips of me singing the song and clips of Miley’s original voice. In both instances, I used AI to switch our voices to match the voice of a celebrity. Initially, I thought the audio would sound exactly like the voices of each celebrity, but I was terribly wrong. It sounds very off, but I think it is worth listening to. I also think it is worth considering when AI will produce audio that sounds exactly like these artists. Already, the option of choice, consent, transparency, credit, and compensation are issues artists are facing with having their work trained by AI models. As AI becomes more advanced, this problem will only grow and the line between infringement and fair use will need to become more clearer. So, give this song a listen, and see if you can identify the voices being used. It is a bit cringy so brace yourself! I will provide the answers at the end.
Because Musicfy didn’t accurately generate the audio to match the celebrities voices, I didn’t include every voice generation in the song. However, I think it could be fun for anyone interested to try and decipher which celebrity you think was used in the following audio clips. I also provide the answers at the end. Goodluck!
Audio 1
Audio 2
Audio 3
Audio 4
Answers to the audio and song!
Link to my reflection paper:
Leave a Reply