> this isn’t that different

It's entirely different. One is fingerprinting ambient audio against a corpus of known signatures to identify music, another is basically silently monitoring and summarizing every conversation it hears for later review.

I don't feel like my privacy is being invaded by a date's phone telling them what song was playing at the restaurant, but I would be disturbed if their phone produced a summary of our conversation, especially if any vulnerable topics came up.

The Android feature is matching songs it hears to an on-device database of fingerprints, the other is recording, transcribing, and summarizing all conversations around it. Seems a little different.