Apple Music and Metadata: No Humans Involved

Update: Apple has corrected this, with a separate artist page for this Genesis. However, the genre is now listed as Christian & Gospel, at least when I view the album on the web. It is still Hip-Hop/Rap in the Music app.


Looking at Apple Music For You this morning, I checked out the New Releases list. I was surprised to see that the latest release in my list was a new recording by Genesis, called No Grey Areas.

“Genesis has a new record,” I wondered? I thought I would have heard about that. I clicked on the album cover, and it said that it was in the Hip-Hop/Rap genre. So I clicked on Genesis to see if there was another artist named Genesis, you know, other than the Peter Gabriel / Phil Collins band, and here’s what I saw:


Oh, so it is the same Genesis? I started playing it…

Nope, it’s a rapper with the same name, and no one at Apple Music could be bothered to make sure that it shows up correctly.

9 thoughts on “Apple Music and Metadata: No Humans Involved

  1. That is quite funny. Apple is not the only culprit though, on Qobuz, if you click on Judy Bailey’s name, the pianist and big band leader from New Zealand, you also get results from the British Born singer. Maybe one day aggregators will realise that they need human workers to check their metadata.

  2. This is only going to get worse. In looking for a song or tune from anyone, especially an obscure performer or band, you can look high and low and get either no results or what appear to be duplicate results but are not, such as this Genesis example. I think—and I may be wrong—that Gracenote is even messed up in the midst of this. Apple Music and maybe Amazon Music have to get the human factor involved, unless they’ve already decided it’s not worth it to pay a staff to sort it out.

  3. I don’t mean to imply that Gracenote is at fault with this. But it does appear that there needs to be a governing entity—almost like a musicologist—to prevent this from happening.

  4. Well, I am sure there are some AI algorithms that could sort this out better before referring to a human. How difficult is it, for example, to check band members and identify discrepancies, or check associated music and discover genre differences. As with Apple Genious and Roon Radio (much better, by the way), recommendation algorithms have to do this all the time.

    • Apple’s Genius works on two levels: one is the actual metadata, and that wouldn’t solve this issue. (With the exception that the genre for this album is listed as Hip Hop / Rap, which should have been a clue.) But Genius also takes into account how many other users have the same music in their libraries, so it can construct a sort of venn diagram to recommend other things. But Genius is only used for your local library; I don’t think that same technology is used on Apple Music more broadly, because there are just too many tracks to make it usable.

  5. I’ve had this happen to me a few times. You know the singer Lobo of “I’d Love You To Want Me” and “Don’t Expect Me To Be Your Friend” fame? His name has come up a few times in my New Music Mix. When I play the songs, they sound so unlike him, like some form of EDM or something. I am also uncertain if this is happening with Enigma (the Michael Cretu named project), as I have checked online and there’s no official news about an album or EP from them.

    I love Genius on iTunes:Music, I use it all the time. I think the only thing that comes close to being similar to Genius is the Similar To song features of Shazam. If Apple were to make Shazam fully Apple Music based, they could incorporate that feature into the main Music app. As they still share Shazam with Spotify, and some Shazam users are Android users, I doubt that will happen too soon.

  6. I think the main reason for this to happen (on tidal is only slghtly better, although with much more poweful info and searches) is that the data must be managed as flat databases extracting data from the metadata of records, instead of a solid, relational database, where ‘Genesis’ would be uniquely identified by an id, as well as the band members, key personnel, tracks, etc. Even some fairly simple AI programming would parse the metadata in relational tables with enough consistency. It is amazing that nobody has done this, but unforgiveable from a company like Apple, not only because of the limitless resources and power they have, but mostly because they once owed their very survival to music.
    Apple should involve seriously on this. It is the least they should to music (and stream in lossless, of couse). But I’ve been waiting for so long for it that I lost every hope and moved to Tidal about a year ago. I couldn’t be happier with the move.
    Btw: all media content by Apple (and most others, to be fair) suffers from this amazing lack of professionalism.

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.