There are a lot of people losing their minds about the discovery of AI-powered music generating tools. For example, there are tools that exist which can be prompted using specific text commands to randomly generate certain styles of music, which just like any other AI generated content, can be further refined. In other words, an AI tool can output something that the user has created just by describing it with words.
More recently however, it became known that it was possible to use artificial intelligence to clone the voice of another artist just by allowing the tool to sample a bit of a specific person's voice. And I'm not just talking about spoken voice, more recently I'm talking about singing voice. Apparently what someone was able to do, was use one tool to strip the existing vocals away from a popular song not performed by Drake (we’ll call it song A), then use artificial intelligence to sample the singing voice of Drake (from any of Drake’s catalog). The AI tool was then able to be prompted to sing the lyrics of song A that had been stripped of the original vocal. The result sounded very close to if not exactly as if Drake had sung the lyrics to that song.
This caused a lot of people to go into a frenzy, the least of which was the RIAA, or the Recording Industry Association of America. They consider this creative act to be a copyright infringement. If you listen carefully, you can almost hear the copyright lawyers salivating at this new development. There's one problem though. The resulting new creation generated by artificial intelligence trying to sound like Drake, but not actually being Drake in real life, would not necessarily be considered copyright infringement, because the content is transformative in a big way.
Even an artist whose voice is being mimicked by an artificial intelligence might feel threatened by the existence of this new tool, but folks, this is just history repeating itself. A similar form of creativity appeared in music culture not that long ago called mashups. That's where clever disc jockeys used the kind of tools that could replace the original vocal from a song with the vocals from a completely different song and mashing the two together. It caused a bit of an uproar, but many of the original artists were amused if not intrigued by this new art form, being inspired to do similar mashup works themselves, while also realizing that these mashups put the spotlight back on these often forgotten songs.
But even if we ignore all of that, this is just a passing fad. It is undoubtedly fascinating to hear a familiar artist sound like they are singing the lyrics of another song. But everyone listening to this knows that it's not real. and the next time an artist puts out something they themselves have created, people will forget about the artificially created content. I mean seriously, which would you rather listen to? A fake generated mashup or brand new, genuine creation from one of your favourite artists?
No comments:
Post a Comment