Paul McCartney & Rosalía: Strategies for surviving AI music

DW (English)

Tens of thousands of AI tracks hit streaming platforms every day. What does that mean for artists – and for us users?

Transcript

00:00Paul McCartney just dropped a new track, but it's not a Beatles-style sing-along.

00:06It's almost silent.

00:09Just tape his and faint studio sounds.

00:11Why?

00:12It's part of a protest album called Is This What We Want?

00:15Created by over 1,000 artists.

00:17The message to tech companies, stop training AI on musicians' work without permission.

00:22Meanwhile, Spanish star Rosalia took another approach to AI for her new album Lux.

00:27She avoided it altogether.

00:29So why are artists pushing back against AI?

00:32And would it really matter to us if music was made by AI instead of humans?

00:37First, AI music is everywhere.

00:40And it's competing with human musicians, especially on streaming platforms.

00:44Tens of thousands of AI tracks are uploaded to streaming platforms every day.

00:49On streaming platform Deezer, it's around 50,000.

00:52That's about one-third of the platform's daily total volume.

00:55And there are even AI stars.

00:57AI country singer Breaking Russ has just landed a No. 1 hit on Billboard with Walk My Walk.

01:05His two biggest hits alone reached 10 million streams on Spotify together so far.

01:10And it's not the first time AI-generated music is charting.

01:13The human producers behind AI-R&B singer Zania Munay even scored a record deal for the project

01:21after Munay had appeared on US charts twice.

01:23The AI-generated voices sound impressive.

01:26But that's exactly what Paul McCartney and other artists are protesting against.

01:31They say AI music platforms steal their creative work.

01:34Zania Munay doesn't just copy one particular singer.

01:38Her voice is built from patterns learned on thousands of real female voices.

01:41And it's tuned to sound appealing to most listeners.

01:44Here's how.

01:45AI music platforms like Suno or Udio work like ChatGPT for music.

01:50You type a prompt and they generate full songs.

01:52They use AI models trained on huge audio datasets often scraped online without artists' consent.

01:59They also learn from user feedback to make tracks that fit popular tastes.

02:04But does it really make a difference to us as listeners?

02:07A recent study has shown that 97% of all participants cannot tell AI music from human-made tracks.

02:14Nevertheless, 51% said they believe the use of AI in music production could lead to generic-sounding music.

02:20And they might be right.

02:22In many cases, AI-made tracks lack the emotional depth of human artistry.

02:26That's not saying that good producers with a vision and prompting skills can't generate AI music that is touching.

02:33But the sheer volume of AI songs can overwhelm platforms and bury quality music in generic content.

02:39Algorithms may favor formulaic AI tracks optimized for engagement and reduce diversity and creativity.

02:46Worst case scenario, AI music platforms force human artists to give up and get stuck in a feedback loop of more of the same.

02:53So, do we need the human touch?

02:56Some artists are hopeful that the human element in music will prevail.

03:00On her recent album Lux, Spanish pop star Rosalia sings in 13 languages.

03:05She could have easily used AI to make it sound authentic.

03:08Actor Adrian Brody did exactly that for enhancing his Hungarian accent in the movie The Brutalist.

03:14And even won an Oscar.

03:16Rosalia considered doing it, but then decided against it.

03:19That's one of the reasons why many news outlets, the Rolling Stone magazine for example,

03:23have described Rosalia's new album as a counterpoint to the AI trend in an era of algorithmic music.

03:29What are your thoughts on this?

03:31Will AI destroy music?

03:32Or will it simply change it?

03:34Let us know.

Category

Transcript

Be the first to comment

Recommended