Counterfeit Drakes and Kanyes: AI-generated hip-hop has suddenly taken over the internet
but is this harmless homage or identity theft?
It is becoming increasingly difficult to determine whether your favorite artist is singing or rapping on tracks that are available on YouTube, Instagram, and TikTok because of powerful new generative AI tools.
In February, superstar DJ David Guetta uploaded a video of himself performing live and playing a song he had written with Eminem's vocals generated by AI, evidently without the rapper's permission. Then there is "Savages," a song by French hip-hop group AllttA that features Jay-Z's AI-generated vocals. According to writer Kyle Chayka of the New Yorker, Jay-Z's "familiar voice adds something ineffably compelling to the track." Recently, Drake and The Weeknd's impressive AI-created collaboration, "Heart on My Sleeve," which some believe is a startup's marketing ploy, went viral on social media.
Every day, additional tracks similar to these are released, many of which are fictitious covers by relatively unknown artists. Also, not every person is satisfied. Keep going week, Drake posted on his Instagram that a computer based intelligence created form of him rapping Ice Flavor's "Chomp (Feelin' U)" was "the straw that broke the camel's back." However, Rihanna has yet to respond to AI-generated versions of her performing songs by Maroon 5, Beyoncé, and Katy Perry, among others.
In the meantime, it is said that Universal Music Group, which has Drake and Rihanna among its many artists, has asked music streaming services like Spotify and Apple Music to collaborate with them to prevent scraping of content that could make music generative AI tools work.
An influential analyst for BNP Paribas Exane recently stated that AI-generated music posed a "new disruptive threat" to the industry's operations, and the music industry may have responded accordingly. There are a lot of legal questions that the technology raises that nobody seems to know the answers to yet.
Fans appear to be enjoying the ambiguity at the moment. Take YeezyBeaver, an Oklahoman 22-year-old whose YouTube channel features Kanye West performing a variety of songs. The most well-known one is a surprisingly charming cover of Plain White T's song "Hey There Delilah." A post on a Ye Discord fan server that linked to a voice model of the rapper and provided instructions on how to insert the model into generative AI tools that can create new music led YeezyBeaver, who requested that BuzzFeed News not use his real name, to begin creating AI-generated music.
YeezyBeaver and a lot of other people use a tool called So-Vits-SVC, which has become so popular that it has a 2 million-view TikTok hashtag. He first tried it by putting Ye's voice on Drake's song "Jungle," an experiment that got a little bit of attention on TikTok. He stated, "I just started looking for songs by other artists that I think Kanye would sound good on." After that, that pretty much got us here.
Ye's voice is also being altered by another user named pieawsome, an American college student who spoke with BuzzFeed News on the condition that his real name not be used. He is a member of the "Kanye unreleased community" on the internet that is dedicated to Ye's unfinished work. Pieawsome said, "I realized we had enough Kanye material to probably make an AI model out of it to finish some of the songs."
So he slashed up a capella segments from the accessible however not formally delivered Ye sound and took care of it into So-Vits-SVC. Following a couple of long periods of preparing, the outcome was great, so he pushed his Kanye voice model out onto Disunity. " Pieawsome stated, "People started working with it, and the produced music got passed around to a lot of people." Soon after, Travis Scott liked an Instagram post with a track Ye made with the model, Ye covering Ice Spice's "Munch."
Jered Chavez, a 19-year-old University of South Florida student, uploaded a mind-boggling AI music video to Instagram: Drake, Ye, and Kendrick Lamar singing "Fukashigi no Karte," the end tune to Scalawag Doesn't Dream of Rabbit Young lady Senpai, an anime series well known on the web. " Chavez stated, "I thought it was a cool concept." It's obvious that I wasn't the first to start it, but my page stands out because I try to add a little humor and a little twist.
According to the student, the comedic approach is intended to protect him from the risks associated with creating the work. With this area of man-made intelligence, there's a great deal of contention and moral worries," Chavez added. " Clearly, individuals that make this music and utilize this computer based intelligence are taking somebody's resemblance and, more often than not without authorization, making something that is basically speaking for individuals."
Chavez pointed out that this could be particularly problematic when generative AI is used to re-create the voices of deceased artists. He stated, "We don't really know what they would want, and they aren't around to give their approval." BohemianRhapsod.ai, which lets users lead a choir of 16 AI-generated versions of Freddie Mercury through perhaps Queen's most well-known song, does not appear to be bothered by this.
No one really knows if hobbyist creators are on shaky legal ground, despite the growing consensus that generative AI could be problematic. Pieawsome stated that he considers his work to be comparable to modding a video game or writing fanfiction based on a well-known book. It's our rendition of that," he said. " That might be beneficial. It could be something terrible. I'm not sure. However, it was kind of inevitable that it would occur.
A different perspective comes from Jonathan Bailey, the former chief technology officer of Soundwide, a music technology company. He stated, "I think you can make a convincing case that using artificial intelligence to reanimate Jay-Z's voice to have him rap or sing something he never created is kind of a form of identity theft."
Bailey continued, "I'm not a lawyer." Legal counselor Donald Passman of Group, Tire, Ramer, Brown and Passman, Inc., who has addressed significant specialists like Adele and Taylor Quick all through his celebrated vocation, declined to say something regarding the utilization of man-made intelligence to emulate craftsmen for this story. He stated that he wished to avoid contradicting a position he might later take in court. Passman commented on the technology, "it's way too new."
It would appear that some record labels have already settled on a decision. Many AI-focused Discord servers have reported copyright takedowns on YouTube channels that post AI-generated music. Chavez stated, "I guess that's one way to deal with it." But I'll be honest: I don't think people will ever stop using this technology now that it's available.
He went on to say, "The judgment of the people who are making [AI-generated music] is the responsibility." I try to make good decisions. Everyone is a little out of their comfort zone here.