Google DeepMind’s AI Pop Star Clone Will Freak You Out
Even if you didn’t watch last weekend’s episode Saturday Night Live, you’ve probably seen it yet. You may already know what I’m talking about: Timothée Chalamet and other similarly dressed actors, shaking the booty in tiny little red underwear. He was, according to the sketch, “an Australian YouTube twink turned indie pop star and model turned HBO actor, Troye Sivan, played by an American actor who can’t do an Australian accent.” » Chalamet and his cohorts were Troye Sivan’s sleep demons, and they haunted straight women everywhere. It was a funny moment and, ironically, the least nightmarish Sivan impression to come out this week.
Thursday, Google DeepMind Lyria announced., which it calls its “most advanced AI music generation model to date” and a pair of “experiments” for music creation. One is a set of AI tools that let people, for example, hum a melody and turn it into a guitar riff, or turn a keyboard solo into a chorus. The other is called Dream Track and allows users to create 30-second YouTube short films using the AI-generated vocals and musical styles of artists like T-Pain, Sia, Demi Lovato and, yes, Sivan almost instantly. All anyone has to do is enter a topic and choose an artist from a carousel, and the tool writes the lyrics, produces the backing track, and sings the song in the style of the selected musician. It’s wild.
My panic about this isn’t fear that a million fake Troy Sivans will haunt my dreams; is that the most creative work should not be so simple, should be difficult. To borrow from A league of its own“It’s Jimmy Dugan, “It’s supposed to be hard. If it wasn’t, everyone would be doing it. It’s the difficulty that makes things great. Yes, asking a machine to make a song about fishing like Charli XCX is amusing (or at least funnew York), but Charli Pop. To borrow again, from a hoisted sign during the Hollywood writers’ strike, “ChatGPT has no childhood trauma.”
Not that these tools are useless. They are primarily intended to help cultivate ideas and, for Dream Track, to “test new ways for artists to connect with their fans”. It’s about creating new experimental noises for YouTube, rather than for the top of the Billboard charts. Like Lovato, who along with other artists allowed DeepMind to use their music for this project, said in a statementAI is disrupting the way artists work and “we need to help shape what that future looks like.”
Google’s latest AI musical toy comes at a tricky time. Generative AI is creating something of a digital copyright minefield, and Google-owned YouTube is trying to deal with both the influx of AI-created music and the fact that it has agreements with labels to pay for artists’ work. appears on the platform. A few months ago, when “Heart on My Sleeve” – an AI-generated song by “Drake” and “The Weeknd” – went viral, it was ultimately removed from several streaming services following complaints from the artist label Universal Music Group.
But even if, say, the manager of Johnny Cash’s estate don’t try to stop AI-generated “Barbie Girl” covers still represent a headache for artists: They can either work with companies like Google to create AI tools using their music, or they can create their own tools (like Holly Herndon and Grimes did), push back and see if copyright law applies to music created from AI models trained on their work, or do nothing. It’s a question that seemingly every artist is thinking about right now, or at least being asked.
