One of the perils of poor metadata is getting your music mixed up
Not so long ago, Google rolled out its music matching service in the US. The feature had been available in Europe already. With the match feature, users don't have to upload thousands of songs to the cloud to listen to their library via Google Music.Any music they have is synced and added to their online library. Technically, they get access to Google's copies of the songs they own.
This means that they get 320kbps streaming for any song, regardless of the bitrate of their original file. Theoretically, a song is a song, it doesn't matter what copy you use, you're going to get the same thing.
In practice, it doesn't always work out that way, as some Google Music users found out, the hard way. Some users report that some of their "explicit" songs are getting replaced with clean versions.
Google is not going all Steve Jobs on people though, as some users are seeing the exact opposite, "clean" versions of songs are getting replaced with explicit ones.
This is just a bug, it's simply a problem of metadata. Clean and explicit versions of the same song differ very little and it's hard for any algorithm to tell them apart, without the proper metadata. Apple experienced the same issues when iTunes Match became available.