帝王会所

Is your new favorite band real? It鈥檚 a bizarre question, but one that music lovers might be asking more often thanks to generative artificial intelligence (AI).

Indie rock band recently went viral after emerging out of nowhere in June and appearing on popular Spotify playlists, according to . With the with band members, as well as an absence of any recorded live performances, many speculated that the group and their music may be the product of AI. The Velvet Sundown amassed nearly a million monthly listeners on Spotify before it was revealed that the 鈥渂and鈥� is not real and their music, is indeed, AI-generated.

Image

帝王会所 School of Media Arts and Studies Director and music industry expert Josh Antonuccio says we are only at the very beginning of how AI is going to transform how content is made and uploaded. One of the most jarring transformations is the increasing use of AI music generators鈥攕oftware tools that use artificial intelligence to create musical compositions in seconds.

鈥淲e鈥檙e in a brave new world,鈥� said Antonuccio. 鈥淲hen you have a [generative AI song] platform like Suno or Udio, the numbers are staggering. They are now generating over 10 songs per second.鈥�

AI鈥檚 ability to quickly churn out content is already extremely visible within the music streaming landscape. On French streaming platform Deezer, , which is nearly twice the number reported four months ago. Other popular platforms like Spotify and Apple Music have shown similar statistics.

AI-generated music is still in a 鈥淲ild West鈥� phase where it is unclear who owns the content and what human-created content can be used to train models. On top of that, there is very little policy that exists to guide these new technologies and processes. This lack of clarity has resulted in turmoil between record labels and AI-driven businesses.

鈥淟abels and publishers understand that these AI models don鈥檛 just make things up, they鈥檙e trained on data,鈥� explained Antonuccio. 鈥淎nd if you鈥檙e training your models on Elton John and The Beatles, you鈥檙e going to get content that mirrors these songs, but you wouldn鈥檛 get that result without using copyrighted work.鈥�

Image

The tension generative AI is creating in the music industry has led to many artists and labels speaking out and even suing AI song generators. In February, a silent album to protest the U.K. government鈥檚 planned changes to copyright law which would make it easier for AI companies to train models using copyrighted work.

鈥淚f you just give copyrighted material to companies to use, they鈥檙e referencing all the human work, all the human capital, all of the intellectual property that these artists have developed, and all of it at no cost. It's insane,鈥� emphasized Antonuccio. 鈥淎nd so that's why you see artists and labels coming out really hard for protections and, more recently, trying to amend some of the decisions in the U.K. surrounding AI. It's not just for them; it's also for how future artists can survive.鈥�

Some of the U.K. artists joined members of the House of Lords in regarding data use and access. Their .

Antonuccio sees the current music industry environment as being reminiscent of the legal battles in the era of Napster and the transition from CDs to digital music and file sharing. Just as agreements were struck that eventually resulted in platforms like iTunes, Spotify and Apple Music emerging more than a decade ago, he believes that labels and publishers will eventually want to license artists鈥� work and come to an agreement with the AI companies.

Image

Some generative AI companies are starting to develop 鈥渇ingerprinting鈥� systems at the point of song generation so that content is marked as such. This development will potentially help streaming platforms, labels and listeners identify and track AI-generated songs. Fingerprinting will not, however, stem the tide of the increasing amount of AI music being created and taking away valuable streaming revenue from humans.

鈥淚 think what we're going to see is more licensing and fingerprinting and watermarking, but that's not going help with the volume of songs getting uploaded and human artists contending with purely AI-generated tracks,鈥� said Antonuccio.

Artificial intelligence isn鈥檛 just generating songs; it鈥檚 also assisting humans and helping accelerate their creative processes and output. There is a big difference between pure AI generation and human artists utilizing AI in their creative processes. With tools like AI-powered lyric generation, mixing, mastering and audio stem stripping, artists have more avenues for music creation than ever.

鈥淭he range of AI tools that are coming, are transforming, and will completely transform the way we think about creating music,鈥� emphasized Antonuccio. 鈥淚f I'm partnering with a tool like Lemonaide, I'm completing songs faster and thus I'm able to create music faster. Something that maybe took me a range of six to 12 months is now down to a range of two to four months. The acceleration process of output is going to grow exponentially.鈥�

Collaborating with artificial intelligence is something that goes beyond the music industry and recontextualizes creativity across the board. Industries like film, animation, advertising, writing and more are openly using AI. Some , not to wholly generate, but accelerate their processes.

Antonuccio says the impact of AI will become even greater as people feel more comfortable using these tools.

鈥淕en Alpha, they're going to be very comfortable using AI to quickly do things like write or ideate or complete lyrics or whatever鈥攖hings that might have taken, previous generations, days or weeks to finish,鈥� said Antonuccio. 鈥淚t really is a transformation period with how we think about what human creativity means and how it's augmented, or how it can be assisted. These tools will add a dimension of possibilities that has been inconceivable until now.鈥�