..And people who spend a lot of time fighting actual spam, as in, actual unsolicited junk email, who have to deal with false positives from uninformed users who think the 'report spam' button is the appropriate response to them getting an Amazon email they don't like.
In practice, there is little difference between "junk email" which I assume you mean scams/phishing/pills/viagra/etc adverts and that Amazon email. Both take space in your inbox, may send you a notification and require time and brain power to deal with. Whether the latter example may comply with some jurisdiction's definition of "spam" is irrelevant.
The "mark as spam" button gives users the ability to keep their inbox clean and you shouldn't be faulting them for using it.
Not to mention, even if we agree for a minute that the report spam button should only be used for emails that conform to the legal definition of spam, which law should we be following? The US' definition of spam is much more liberal than the EU GDPR's one for example.
Creating an account with a company is not signing up for mailing lists unless there's a choice presented. Yeah, you can hide consent in the ToS that nobody reads, but that's not asking for permission.
It's not a false positive. The filter needs to be tuned to what your users think is spam, that is what spam filters are for. You are not the gatekeeper of what other people are allowed to think is spam.
Your opinion isn’t common among anyone other than marketers trying to justify sending spam.