- YouTube Music users say AI slop is filling their feeds
- AI-generated fakes are a growing problem within music streaming
- AI can be a creative tool, but slop is just spam
YouTube Music subscribers are sending Google a clear message: stop the slop!
As spotted by Plunikaweb, the same AI slop that’s been a problem on Spotify now appears to be affecting YouTube Music too, with people on the YouTubeMusic subreddit venting their frustration.
The conversation was started by user vlastawa, who wrote that on opening YouTube Music “six out of ten News recommendations were AI slop. The other day every other song in my auto-generated playlist was AI slop. Clicking ‘I’m not interested’ or thumbing them down makes no difference as it only applies to this specific song, not the ‘artist’… dear Google, that’s not what I’m paying for.”
As one commenter noted, it can be quite easy to spot the slop: “If there are 545 albums released within one year [by a single artist] that’s a good sign”.
So what’s going on, and why are people – including me, a recording musician – annoyed?
Why streamers need to stop the slop
The problem is really simple: AI slop is spam, whether it’s AI music, AI social media posts or the 300 AI-generated fake-bookclub scams that hit my inbox every week. And for customers of services such as Spotify and YouTube Music, it’s spam they don’t want in their music feeds.
And it’s worse than just spam. When I see AI slop on a streaming site, I know that said slop has almost certainly been based on stealing from artists: the data the AI uses to create soundalikes has in part been based on illegally harvested music. In some cases it’s also impersonating the work of actual artists with fake songs appearing on their official pages. And most artists are not Metallica-scale megastars who needn’t worry about the odd rip-off: AI will happily copy struggling or up and coming new artists too.
AI slop in streaming is a real worry for many musicians I know. The more AI slop gets recommended and streamed, the less music by real musicians gets streamed and the harder it becomes for them to get noticed (or paid) – so artists who are already struggling to make money stand to make even less than they’re already getting.
This isn’t a knee-jerk, anti-tech thing. There’s a huge difference between musicians in the 1970s being snobby about synthesizers or damning drum machines and people being upset about AI slop filling their feeds. AI can do brilliant things in music, from creating realistic drums to helping you make your song sound bigger, and musicians have been experimenting with generative music for decades now with often great results. But slop is just slop. It’s music made to game the algorithm, not connect with an audience. It’s not art to be listened to but content to be snuck into your feed when you’re not paying attention.
The issue isn’t grumpy old musical purists yelling at kids to get off their musical lawn: it’s an objection to streaming platforms allowing (or even possibly encouraging) actual music to be replaced with Muzak, with pale imitations of real artists being pushed into discovery and For You feeds.
It’s not as if there’s a shortage of new music either – and every tossed-off bit of AI slop pushed into your recommendations means one less opportunity for that music to be heard, for you to hear someone who might be your next favourite artist, and for that artist to be able to make a living from what they do.
The best music has heart and soul and fire in its belly, and the worst is just a half-baked copy of what someone else has already done. AI tools can and will help artists make more of the former, but AI slop just regurgitates ever more of the latter.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.
https://cdn.mos.cms.futurecdn.net/iAkcihZQCbuLB66zMuJ9ef-1920-80.jpg
Source link
