Would you know if a track created entirely by computer code popped up on your playlist?
Probably not, given that when 9,000 people were asked if they could identify a track made by AI versus one made by humans, 97% of them failed to do so.
Music created by AI is going mainstream, but most streaming platforms still aren’t letting us know if we’re listening to – and sending royalties to – an actual person.
Take Xania Monet, the first AI-generated artist to break into the US Billboard charts. Her top song has now been streamed over six million times, but there’s no indication she’s not a real person if you look at her Spotify profile.
So if streaming platforms aren’t going to tell us, can we tell from listening?
Listen to the tracks below, and see if you can detect which are made by AI. It’s at least one, and could be both. Scroll to see the answers below…
French streaming platform Deezer is the only mainstream music platform to proactively tag AI-generated music.
They conducted the survey alongside Ipsos, asking people in eight countries including the UK, US, Brazil, and Japan, if they could identify AI songs.
Results were overwhelmingly in one direction: 97% of people couldn’t tell the difference.
But they wanted to know, with 73% saying streaming platforms should be upfront if they are recommending AI music, and 70% saying that fully AI-generated music is a threat to the liveihoods of musians,
Over half (52%) of people said fully AI songs should not even be included in the mainstream charts.
It’s now shockingly easy to generate slick music in whatever genre you want, with apps like Suno and Udio.
Where AI image and video generators like Soro and Nano Banana have gone viral, music hasn’t had as much same attention, which is perhaps why streaming platforms can get away without the same signposting that we demand on visual apps like Instagram.
Amber Mark, an American singer songwriter, said she is ‘comforted to know that people care about music and value real songs over AI generated content’.
Concerned about what the rise in AI music means for creators, she added: ‘The fact that such a large majority of people want proper safeguards implemented around AI music sends a strong message to the industry that it’s time for action.
‘I’m also happy to see that so many are on the artists’ side when it comes to making sure that AI-models can’t train on music without the consent of the creators.’
So which tracks were AI generated?
Ashes on the Carousel by the Velvet Sundown is entirely AI generated, and was used in the survey.
The ‘band’ previously caused controversy by racking up millions of plays on Spotify, with their profile claiming it was made up of real musicians.
But the other track, Pretty Woman by Juicy Lucy, is all human-made.
Did you get it right?
-
I already knew the bands so not a fair assessment
Deezer said they continue to be flooded with AI tracks, which now number 50,000 a day, 34% of their total intake.
Many will be submitted to other platforms too, often in a bid to make some easy money from royalties if they are listened to enough.
Spotify has announced steps towards being more transparent, but still does not tag AI generated music, previously saying it’s challenging to have a binary tagging system due to some legitimate artists using AI as part of their creative process.
But this leads to a situation where people could be listening to an entirely fictitious pop superstar with no idea, such as Xania Monet whose artist profile is blank and gives no indication of the truth.
Critics say computer generated music raises ethical questions about copyright on what the tracks were trained on, as well as threatening the livelihoods of musicians.
As Ai music goes mainstream, it’s not only photos and videos we need to ask questions about.
Get in touch with our news team by emailing us at [email protected].
For more stories like this, check our news page.
MORE: Matthew McConaughey and Michael Caine slammed for ‘selling their souls to the devil’
MORE: Mum-of-two says dentist called police during appointment over use of AI
MORE: I thought I’d seen it all – then my pupil brought up chemtrails
