The controversy around AI slop in the best music streaming services is still a huge topic of conversation, and recently Spotify has borne the brunt of user backlash.
There’s no doubt that fully AI-generated music is making the rounds and more users are coming face-to-face with it, but as these AI models get smarter, it becomes more difficult to tell human- and AI-generated music apart.
“AI music and traditional music are becoming indistinguishable. Trying to separate the two today is a bit like trying to avoid music that used synthesizers in the early 1970s. The technology is part of the music.”, says Dr Nicolai Klemke, Founder and CEO of Neural Frames, an AI animation generator.
That said, there’s still a plethora of ways for you spot when AI music makes an appearance in your music streaming app and with the help of industry experts, here are the three main ways to spot AI-generated music – and how you can report them in Spotify.
How to spot AI-generated music in Spotify
1. Check discography
One of the main ways to spot AI music in Spotify is by reviewing the artist’s discography, and not just taking a glance at their album art. If an artist’s discography dates back only a year or two, it often raises suspicion of them being AI-generated, although it could always be a non-AI artist who has simply emerged within a short timeframe.
If this is the case, reviewing the artist’s music output pattern is another good indicator. Normal artists always abide by a release structure, giving themselves ample amounts of time in between each project drop, but AI-generated artists can get away with churning out music at whatever rate their creators wish.
For example, the popular AI artist Sienna Rose released over 45 songs between September 28 and December 5 last year according to the BBC, an amount that’s unfeasible even for an artist like Taylor Swift, who’s notorious for pumping out albums annually.
This is a characteristic of AI-generated music that industry experts have also picked up on, including SEO Director at Paperstack Caleb Johnstone, whose expertise has provided him with insight into how to identify when synthetic music is being used to flood streaming platforms. He provided the following comment:
“I have noticed that the virtual profiles of AI-generated artists do not produce any digital footprints and that they will upload music at an incredible rate, for example, uploading 50 tracks in two months without any type of promotional activity or fanfare. I have never seen a legitimate musician produce music at this rate.”
Social media and music release go hand in hand, and it’s near impossible for a real artist to have a career without a strong social media presence to promote their music. The next place you should check for red flags is the artist’s metadata consistency, and if an artist on Spotify has little to no social media presence, it’s likely they could be AI-generated.
Aside from announcing new music, artists use social media to document everything from live performances, press interviews, and other announcements such as live tours to promote their music. If these types of content are missing, it’s a red flag.
“Authentic composers regularly share studio gatherings, live conducts, or private tales that show their journey. By cross-assign to source these visual and audio entertainment transmitted via radio waves cues accompanying the expert’s connected to the internet occupancy, I can frequently establish either music is AI-created or really human-created,” says Philip Heusser, President and Co-Founder at Motif Motion animation studio.
Though there are some AI-generated artists with social media (Sienna Rose is somewhat active online), there are other signs that point to them being a product of AI such as not being verified artists on streaming platforms, low follower counts, and social media feeds that aren’t as tailored or aesthetically pleasing as legitimate or higher-profile artists.
3. Listen out for ‘generic’ sounds
The common ‘telltale hiss’ heard in AI-generated music is one way to tell a composition apart from one that a human has produced, but this is becoming increasingly difficult to distinguish as AI models get smarter. That said, there are other red flags to look out for in the audio.
Examining the way specific instruments sound is one way to tell if a song has been AI-generated, which is what scientist turned artist Psients suggests: “Listen to the ‘crack’ of a snare drum or the pluck of a guitar. In AI music, these sounds often feel ‘soft’ or ‘pillowy,’ lacking the sharp physical impact transients of a real instrument.”
Unlike the dynamic sound of music made by a legitimate artist, AI music often sounds ‘generic’ because of repeated and unadventurous beats. Paul DeMott, Chief Technology Officer for Helium SEO, offers further advice on this: “Listen for repetitive melodic patterns and predictable chord progressions in different songs. AI models recycle common patterns because they optimize for familiarity. If you have multiple tracks with almost identical hooks or drum patterns with slight differences, then music very likely was generated using templates. Human creativity results in greater variance.”
How to report AI-generated music in Spotify
If you think you’ve run into an AI-generated song on Spotify, the platform has resources you can use to flag it for both subscribers and artists.
Subscribers can report AI music on Spotify’s Safety and Privacy Center. Though there’s no ‘report AI’ option, you can select the option for ‘deceptive content’. Similarly, artists can also report suspected impersonation through the Spotify content mismatch process.
Alternatively, there are third party sites you can use to check if a song is AI-generated including DeepMatch, letssubmit.com, and Find AI Voice, however most require you to upload MP3 files of songs.
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/fE5drui3e3FPE9ab4GjkHJ-594-80.jpg
Source link
rowan.davies@futurenet.com (Rowan Davies)
