Harmonies Unleashed: Exploring the Captivating World of AI-Generated Music  | by Gurkeerat Kaur | Medium

Much of the conversation around AI-generated songs rightly focuses on copyright and artist rights, but other issues remain to be resolved, such as the quality of the music. Michael Sayman worked at Facebook, Google, Roblox, and Twitter. At 26 years old, this software engineer already has a book about his memoirs, App Kid. But until he started developing his latest project, he had never created a website. “I did it in five hours over the weekend, out of frustration trying to make sure there wasn’t anything like it,” he says. “Now the page has almost a million streams .”

Harmonies Unleashed: Exploring the Captivating World of AI-Generated Music  | by Gurkeerat Kaur | Medium

Sayman’s site is AI Hits. Since its release in April, he’s been collecting a controversial new musical trend: songs created with artificial intelligence tools that imitate, with chilling precision, mainstream music stars like Drake and Kanye West. The conversation about AI-generated music has been frenetic and it has been plagued by fears about what it means and what it portends. But Sayman is a technology optimist. That’s why he created the Hot 100 list for songs made with artificial intelligence.

Great musical hits intervened with artificial intelligence

AI Hits sifts through the growing musical waste and ranks songs based on their aggregate plays across the various platforms they’ve been released on, linking directly to the tracks, unless, of course, they’ve already been taken down by the time you click. With its remarkable fidelity to real music, the Drake-simulating ‘ Heart on My Sleeve ‘ became the first ‘hit’ of the AI ​​music era, and various versions of it dominate AI Hits. Sayman points out that there is even one with AI-generated voices of Joe Biden. Covers by Ariana Grande, Travis Scott, Juice WRLD, SZA, and Lana Del Rey also appear on the list.

Sayman uses the term ‘voice’ to refer to the imitated artist and uses ‘artist’ to refer to the username of the person who created the song. This use of nomenclature may seem trivial, but it is significant. It is a step towards creating a shared lexicon around all of this. The landscape of AI-generated music is endless and messy on a discursive level, but as Sayman points out, we are all participating in the beginning of a conversation that will unfold over the years. “How do you search? Who are the creators? How are brands assigned to them? How are the revenues shared? And how does it work when you can make hundreds of remixes of the same song?

This last question, about the legality of the practice of music generated by artificial intelligence, is fundamental. Spotify quickly removed ‘ Heart on My Sleeve ‘ and the record label that represents Drake, UMG, pressured the company to remove thousands of other songs created with AI. In a recent podcast interview, Ice Cube urged Drake to directly sue the creator of ‘ Heart on My Sleeve ‘ and tweeted that he found the idea of ​​generating a song in the style of a deceased artist without the approval of his estate ” evil and demonic.” But beyond the possible legal or ethical blunders of AI, others, from avant-garde musicians like Holly Herndon to established artists like the Pet Shop Boys, are supporters of AI as a creative tool. It could even give rise to a new musical genre.

Sayman believes that artificial intelligence can lead to a more democratic and open music industry. “Record labels used to have all the power: they were in charge of distribution, resources, and production quality. Social networks replaced music distribution and discovery. Today we see AI expanding production quality so that there are opportunities for more people to participate in the music creation process. More Drake singles! Instead of having two or three producers, he can count on millions working on those songs!” He laughs. “I’m only ‘half’ joking,” he added.

Sayman may be speaking without knowledge of the facts, but futuristic pop star Grimes has already embraced the idea of allowing an infinite number of non-professional musicians to make songs with her voice. Through a site called Elf. tech, he makes his voice available to anyone for commercial use, in exchange for 50% of the copyright. In a recent interview with The New York Times, he reviewed the music already created through the website, and his heartfelt embrace of the songs provided a refreshing and much-needed counterweight to the hysteria around AI. “What I like about early AI pieces is that you can hear the technology very deeply,” he shared. “People will appreciate it more in five years when they realize that people did things like that for just a couple of months.”

The role of record companies in the face of AI in music

‘ Heart on My Sleeve ‘ was met by audiences in part with almost visceral revulsion, and it’s possible that reaction was rooted in horror at how easy it seemed to create a fake Drake. But as news editor H. Drew Blackburn noted in an article for Bloomberg, “Drake has been making music that sounds like an AI Drake for years.” People are collectively obsessed with the theoretical ramifications of AI-created music. But maybe there’s another question everyone forgets to ask about all this stuff: Is it any good?

Marc Weidenbaum is a writer, sound artist, and professor at the Academy of Art University in San Francisco (USA). “I find it strange that there is so much talk about music made with artificial intelligence,” he says. “We have been creating artificial life since at least the Golem (the time of the Jewish ancestors).”

Ultimately, for Weidenbaum, something like ‘ Heart on My Sleeve ‘ is negative, not because of its apocalyptic connotations, but because it is, well, boring. In his opinion, it is something similar to cybernetic music, in which an artist programs a machine to explore and create sounds that are beyond the author’s control. Random or improvisational music works similarly. “Unintended consequences are a feature, not a flaw,” he says. “Brian Eno was interested in the idea of ​​the composition being a garden, constantly changing.”

But using artificial intelligence to simulate Drake is very simple. “Most people don’t create good art by copying and pasting,” Weidenbaum says. “What makes pop work is that it always changes and reacts. It is nothing more than a feedback loop between systems.”

For Joey DeFrancesco, a musician and organizer with the Union of Musicians and Allied Workers, what’s alarming about artificial intelligence is how it can ultimately be used by major record labels. “Any potentially interesting artistic use of AI pales in comparison to the corporate domination of such technology that will inevitably occur under our current music industry power structures,” DeFrancesco states. “The fantasy of technological capitalism” in the music sector “has always been to completely eliminate artists and end the need to pay royalties.”

DeFrancesco points out that artists have already successfully fought against the harms caused by new technologies in the past: “In the 1940s, musicians went on strike to demand that the profits generated by the new technologies of vinyl records be shared with the musicians, and they won. He also refers to the current Writers Guild of America strike: “The big studios want freedom of action on artificial intelligence so they can appropriate the work of writers and then do without them entirely. But the screenwriters say ‘no’ collectively “. And DeFrancesco has reason to worry. After making a fuss over ‘ Heart on My Sleeve ‘, UMG was quick to embrace machine learning technology, partnering with an artificial intelligence company called Endel.

All of this reminds us that theorizing about the worst possible scenarios ignores the already bad ones that musicians and fans face today. The pressing questions about AI-powered music are human. Does its existence mean that musicians will be fooled in new ways? Is it worth listening to it? We still don’t know the answer to either.

Sayman is currently working to improve the search function and address requests from passionate AI Hits users. As for Sayman’s listening habits, her favorite song is ‘Por qué’, a duet between Rihanna and Bad Bunny. Sayman speaks Spanish and her family is Peruvian. “It’s a Spanglish song,” she says, “but it’s funny: Spanish doesn’t sound too good.” The artificial intelligence that created the song was trained on a data set that did not contain enough Spanish. Series. It doesn’t matter. For him, it is still a great success.

By win11

Leave a Reply

Your email address will not be published. Required fields are marked *