When pop groups and rock bands train or play, they rely on their guitars, keyboards and drumsticks to generate music. British AI musician Oliver McCann, better known by his stage moniker imoliver, activates his chatbot.
The genres of McCann’s songs are diverse and include country rap, electro-soul, and indie-pop. There is only one significant distinction between McCann and conventional musicians.
He claimed to have no musical ability at all. He has no musical training and is incapable of singing or playing an instrument.
With experience as a graphic designer, McCann, 37, began experimenting with AI to see whether it might help him be more creative and “bring some of my lyrics to life.” After one of his songs received three million plays, he signed with independent record company Hallwood Media last month. This is reportedly the first time a music firm has signed an AI music creator to a contract.
McCann exemplifies the rise of synthetic music caused by ChatGPT-style AI song generating tools such as Suno and Udio. The movement was most prominently emphasized by a fake band called Velvet Sundown, which went viral despite the fact that all of its lyrics, music, and album covers were produced by artificial intelligence.
It stoked discussion about AI’s place in music while igniting concerns about “AI slop”—automatically created, mass-produced, low-quality output. AI song generators, which are democratizing songwriting but also posing a danger to the music business, were also brought to light.
Music experts predict that generative AI will revolutionize the industry. Only a few facts are available, though, regarding its effects on the $29.6 billion worldwide recorded music business, which includes around $20 billion from streaming.
Although they only make up a small portion of all streams, the music streaming site Deezer provides the most accurate statistics, estimating that 18% of the songs posted to its platform each day are entirely artificial intelligence (AI) created, suggesting that relatively few people are actually listening. Spotify and other larger streaming services have not disclosed any data about AI music.
Udio denied to provide information about its user base and music production volume. An attempt to reach Suno for comment was not answered. In addition to pro and premium tiers that grant access to more sophisticated AI models, both offer free basic levels.
“It is a total boom. Josh Antonuccio, the head of Ohio University’s School of Media Arts and Studies, described it as a “tsunami.” As young people grow up with AI and get more accustomed to it, the quantity of music produced by AI “is just going to only exponentially increase,” he added.
With its capacity to produce seemingly original content, generative AI has caused a division in the music business, with musicians and trade associations protesting that recorded music is being used to train AI models that drive song creation tools.
While recording artists fear that AI music startups may diminish their originality, record companies are attempting to offset the danger to their revenue sources.
For copyright infringement, Suno and Udio were sued by three major record labels last year: Warner Records, Universal Music Group, and Sony Music Entertainment. According to reports, the parties also started talks in June that may go beyond resolving the litigation and establish guidelines for how musicians are compensated when AI remixes their compositions.
The German royalty collecting group GEMA has filed a lawsuit against Suno, claiming that the company is creating music that is identical to Lou Bega’s “Mambo No. 5” and Alphaville’s “Forever Young.”
In protest of proposed changes to U.K. rules on AI that they believe will undermine their creative autonomy, over 1,000 musicians, including Kate Bush, Annie Lennox, and Damon Albarn, produced a silent album. In the meanwhile, other musicians have embraced the approach, including Timbaland, Imogen Heap, and will.i.am.
Some users claim that the discussion is only a repetition of previous disputes about once-novel technology that later gained widespread usage, such synthesizers, drum machines, and AutoTune.
People gripe that “you’re doing all the work for you with a computer.” That is not how I perceive things. I see it like any other tool that we have,” Scott Smith, whose AI band Pulse Empire was influenced by British synthesizer-driven bands such as Depeche Mode and New Order in the 1980s, said.
In Portland, Oregon, Smith, 56, a semi-retired former public relations officer in the U.S. Navy, stated that “music producers have lots of tools in their arsenal” to improve records that consumers aren’t aware of.
Smith never learned to play an instrument, just like McCann. They both agree they spend a lot of time and energy creating their songs.
Smith writes the lyrics in only ten minutes when he gets inspired. He will then create many versions for up to eight or nine hours until the tune “matches his vision.”
By prompting and re-prompting the AI system, McCann said he frequently produces up to 100 distinct renditions of a song before he is pleased.
Although AI song generators are capable of producing both music and lyrics, many seasoned users would rather compose their own lyrics.
“AI lyrics tend to sound pretty dull and cliche,” McCann stated.
Although AI lyrics are typically “extra corny” and less imaginative than human lyrics, Lukas Rams, a Philadelphia-area resident who composes music for his AI band Sleeping With Wolves, claimed they can benefit at the beginning of the composing process.
Rams, who writes his own lines, stated, “It will perform very simple rhyme schemes and it will repeatedly repeat the same structure, sometimes while he is putting his children to bed and waiting for them to go to sleep.” “And then you’ll see words like ‘neon,’ or anything with’shadows,’ which are very indicative of lyrics created by artificial intelligence.”
Before employment and family obligations began to take up more of his time, Rams played drums in high school bands and worked with his brother on original tunes.
After that, he learned about AI, which he utilized to produce three Sleeping With Wolves albums. Making a CD jewel box with album art shows that he has been taking it seriously. His tracks blend EDM with metalcore, and he intends to share them more broadly online.
“Just to get it out there, I do want to start posting this on YouTube, social media, distribution, or whatever,” Rams said. “If I don’t, I might as well be the only one who hears this stuff.”
According to experts, AI has the potential to revolutionize the music industry’s production process by enabling anybody to write a successful song.
“Just think about what it used to cost to make a hit or make something that breaks,” Antonuccio continued. And that simply continues narrowing down to a bedroom, a laptop, and a large studio. Now, it resembles a text prompt—multiple text prompts.
However, because to the lack of legal certainty on copyright, he noted, AI music is still in its “Wild West” phase. The legal fights against file-sharing websites like Napster more than 20 years ago, which signaled the shift from CDs to digital media and ultimately cleared the path for today’s music streaming services, are comparable to this, he said.
AI will ultimately join the mainstream music industry, according to creators.
“I believe we are entering a world in which the next big hit could be made by anyone, anywhere,” McCann remarked. “I think it opens up the possibility for AI music to be featured in charts as AI becomes more widely accepted among people as a musical art form.”
