AI has arrived through music. Until now, there is a big sea between people and algorithms

The production is slick, the drums are on point and the vocals sound great, but a titan of the Newfoundland and Labrador music scene hears something strange about “It Could Be Worse” and “Tales of The Atlantic,” two songs generated in less than of one minute using a powerful algorithm.

“He’s a country singer, so that’s wrong. And the lyrics don’t really rhyme,” said Bob Hallett, a founding member of the Newfoundland folk-rock band Great Big Sea. “It just sounds a little strange.”

Hallett had just finished listening to the fun tunes, which were created using a generative AI tool called Suno using cues that could describe any Great Big Sea jam: Celtic, folky, upbeat, passionate.

They missed the mark, Hallett said. On a scale of one to 10, with 10 being the band’s hit “Ordinary Day,” he gave the songs a two.

But experts say that rating could rise quickly. Technology like Suno’s is advancing rapidly and its performance is only improving, said Jimmy Lin, professor and director of the artificial intelligence institute at the University of Waterloo.

Suno is one of several companies creating generative AI software that allows users to create original songs using text prompts. People can create instrumental tracks or songs with lyrics, which can be generated by the program or provided by the user. But if the user provides copyrighted lyrics (the first few lines of Nirvana’s “Smells Like Teen Spirit,” for example), the program won’t create the song.

Suno also doesn’t make songs meant to sound like other artists’ work, such as “Heart on My Sleeve,” the AI-generated song using unauthorized vocals by Drake and the Weeknd that caused controversy in the music industry last year. . When asked to make “a Great Big Sea song about cod fishing,” Suno’s resulting melodies had a pensive, Celtic style, but they didn’t sound at all like an authentic song from the band.

Google is working on similar software, called MusicFX, which can be tested through its AI Test Kitchen site. And Adobe introduced Project Music GenAI Control last month, which it described as an “early-stage generative AI music generation and editing tool.”

In December, Microsoft introduced a Suno-powered song generator for its Copilot chatbot, which is a program that uses artificial intelligence to simulate conversation with users.

The technology behind these programs is similar to that which powers the ChatGPT chatbot.

Lin said such programs use massive data sets to “train” algorithms, or step-by-step processes, to take any starting point and predict what will be next. So while chatbots trained with text can predict the next word in a written response, a music-generating program is trained using sound to predict the next “acoustic sequence,” he said.

The New York Times sued Microsoft and ChatGPT owner OpenAI in December for using its stories to train programs.

Lin said the companies behind AI platforms that create music could find themselves in similar trouble if they have trained their algorithms on the work of artists who have neither consented nor received compensation for their music being used in that way.

“It’s an unresolved question whether this is fair use or not,” he said in an interview. “It will take care of itself. The court always does.”

Suno’s website does not indicate what data it used to create its program, and the company did not respond to a request for comment.

Hallett said he wouldn’t be surprised if the algorithms had learned from some Newfoundland bands. He said the songs he produced had some characteristic sonic trademarks, including tight melodies and heavy strumming on acoustic guitars, that he and his fellow producers have cultivated over the years recording albums for bands such as Ennis Sisters, Shanneyganock and The Fables.

However, he was unfazed by the music used to train these programs, noting that artists have long been competing with platforms like YouTube and Spotify that already deeply reduce musicians’ earnings.

“There’s a feeling of surrender in all of this… it’s very difficult to control,” Hallett said. “Creative work is really about driving concert sales or finding commercial locations. Even at the highest levels, people don’t make money selling records anymore.”

Lin said the AI-generated songs will likely be used by advertisers who need a catchy jingle for a commercial. And she believes that could start happening soon given the rapid pace at which these tools are evolving.

“We’re not talking about years or decades. We’re talking about months,” Lin said.

But Hallett said anyone looking to use music to connect with an audience would be better served by humans.

“It’s easy to be afraid of AI,” he said. “But we’re all drawn to sincerity in music. We want to hear people telling us a real story and giving us real emotions. And the computer just can’t do that.”


This report by The Canadian Press was first published March 3, 2024.


With files from The Associated Press

Leave a Comment