| In early May, Grimes, the elven, extremely online singer and tech enthusiast, made headlines by announcing that she'd allow anyone to use her voice to make a song using generative AI. And not only that – she'd split any resulting royalties 50/50. At the time, the story offered a counterweight to the growing hoopla surrounding Ghostwriter977, a mysterious online creator who had recently published a song seemingly featuring the synthetic voices of Drake and The Weeknd that gained millions of streams across platforms. That track landed around the same time Universal Music Group, to which Drake and The Weeknd are signed, began its counter-offensive against AI-generated songs, particularly ones made without proper licensing. Pundits gnashed their teeth about the forthcoming onslaught of these software-created tunes and bemoaned a future where anyone could capitalize on a prominent artist's voice or style. Could anyone become a hitmaker on the backs of established musicians? It's now been three months since Grimes launched her generative AI software, Elf.tech, and in that time, more than 300 songs featuring GrimesAI have been submitted for distribution, according to the New York Times. Many have made it onto Spotify and started accruing royalties. "Feeling really amazing from making beautiful art is something that has typically been behind a gate for a lot of people — extreme amounts of time and energy, years of technical training," Grimes told the Times. "I think it's valuable that there's a tool with which, if you have a beautiful idea, you can make a beautiful thing and access that." And despite the prior sky-is-falling narrative about what would happen if fans could access the voices of their favorite artists — namely, that they might rapidly displace them — none of these new tracks have achieved viral status or become certifiable hits. The most popular song, Cold Touch, is at 852,210 streams since its release in mid-May, shortly after the tech debuted. That's only around 300,000 streams more than Grimes' own latest single, I Wanna Be Software, which dropped last Friday. No other GrimesAI tracks have cracked 100,000 streams. The early results of the Elf.tech experiment suggest that Ghostwriter977 might be an outlier rather than the rule. At least for me, knowing Grimes co-wrote I Wanna Be Software piques my listening interest more than if a fan put together the same track. Grimes is saying she wants to be software! Think of the culture! The growing Grimes lore! Of course, the lackluster commercial success doesn't mean Grimes' AI tech is a failure. It's still early in the generative AI world, and it's important to understand how fans might use officially sanctioned tools. (Users can submit their tracks to Grimes' team for sign-off and to be properly registered and distributed to the streaming services. They also might be featured on the official GrimesAI Spotify playlist by doing so.) Cherie Hu, founder and publisher of Water & Music, a research and intelligence network, worked alongside producer and writer Yung Spielburg, or Lenny Skolnik, to publish their own GrimesAI track, Eggroll. Ultimately, they both found gratification just knowing that Grimes' team heard their track. "When we submitted our song to the hub, I really hoped Grimes might hear it and love it and share it," Yung Spielburg wrote. "There was then huge satisfaction from both myself and Cherie hearing positive feedback back from Grimes' team." In a separate text thread with me, Hu said that shortly after their song dropped, they began working with Thrice Cooked Media, an indie artist management and label services fund, to build out a "long-term strategy" around marketing the song. "I'd say our main purpose in putting out the song was to be an early mover in engaging with the latest music AI tech to sharpen our understanding of where it might be headed, in a way that also meaningfully engaged with the community around an artist as cutting-edge as Grimes — and, most importantly, felt fun," she wrote. "Any subsequent virality is icing on the cake." Elf.tech doesn't address labels' very real concerns over tech companies training their models on the hard work and creativity of artists, beyond the more identifiable, direct use of their voices. But it does give us a window into how labels might fare if they do empower fans to produce their own work using a star musician's voice. So far, the threat appears minimal. Starting on July 24, TikTok began inviting podcasters to put their shows on the platform. Podcaster and producer Marcus dePaula spotted the invite and posted about it on X, the platform formerly known as Twitter. He walked me through his test of the feature, which requires podcasters to submit their RSS feeds to TikTok. Once they do so, they can select an episode of their show and link it to a specific video. That video will then have a tiny banner inserted at the bottom that says "Listen to the podcast." Tapping through loads a TikTok audio player that continues playing even if a listener locks their phone. A TikTok spokesperson confirmed the test of RSS feed integration in the app and said the platform has formal agreements with several hosting providers to work with them. So TikTok now has a full-blown audio player built into the app, at least for some people. This could end up becoming a worthwhile feature for podcasters looking for new ways to promote their shows. We've previously discussed some of the more innovative ways podcasts show up on TikTok specifically, so if clips could actually result in a listener directly hearing a full-length podcast, that'd be a win. However, as of right now, TikTok doesn't have a backend analytics system that accounts for podcasts, and is serving up the shows itself, meaning podcasters won't have data about how many plays they're receiving. (The spokesperson said the platform would look to introduce new features in the future, and presumably, something related to analytics would be on that list.) For now, though, we can say it at least seems a new podcast player has potentially entered the arena. Keep me posted, podcasters, as you play around with the tool. | 
No comments:
Post a Comment