We are living in interesting times. Artificial Intelligence is making strides; programs like ChatGPT and Bard have been destabilizing disciplines. ChatGPT is making it harder for teachers to know if the paper they are reading was written by one of their students. Midjourney is making it hard for a lot of people to know when the Pope is and isn’t wearing an extremely stylish puffer jacket.
But works produced by these programs aren’t indistinguishable from human works yet. With AI pictures, you can look to the teeth and hands to spot tech art. At least that used to be the case. With AI “authored” writings, there are a couple of anti-plagiarism programs you can feed the writing through. But what are the tells for AI generated sounds? The music industry is going to need to figure that out. And quickly.
In a letter to streamers including Spotify and Apple Music, the record label Universal Music Group expressed fears that AI labs would scrape millions of tracks to use as training data for their models and copycat versions of pop stars.
UMG instructed the platforms to block those downloads, saying it would “not hesitate to take steps to protect our rights and those of our artists”.
If I heard this, I’d think that Kanye West’s PR found a way to get him back in good graces and SNL was in desperate need to figure out what they were airing next week, not an algorithmic or computational hodgepodge meant to trick me into thinking that what I was listening to was a genuine expression of human creativity. Ok, who am I kidding? The music industry hasn’t been that for a while now. Formulaic songs have been destined for the Pop 100 for a while, but humans were at the helm of the ship and reaping the rewards of their intellectual property. And we aren’t too far from AI being the captain now.
OpenAI’s Jukebox has been used to generate songs in the style of Katy Perry, Elvis and Frank Sinatra, while an AI-generated Jay-Z was so good it sparked one of the first successful copyright strikes, after the artist’s agent, Roc Nation, got the song pulled from YouTube.
Other systems, like one demonstrated in a research paper by Google, are capable of generating entirely new compositions from text prompts such as: “Slow tempo, bass-and-drums-led reggae song. Sustained electric guitar. High-pitched bongos with ringing tones. Vocals are relaxed with a laid-back feel, very expressive.”
This bears emphasizing. Not only are there dupes being created that are close enough to trigger copyright strikes, there are systems able to create novel pieces.
You should prepare to see lawsuits attempting to preserve musician’s IP. And for an influx of songs that aren’t being performed by the people you usually associate with performing them. Try to see if you can hear the extra fingers and teeth.