Again, that's not how spam filters actually work in the real world. They hardly care about message contents. They care about how many spam reports users file and those are weighed against non-reported messages for any stable identity like phone number, sending domain name, etc.

Spam meanwhile is defined as messages users don't want. If users can't tell the difference between spam and non-spam then, by definition, the messages are not spam.

If you send lots of "legit" messages to people who don't want them you'll be reported as a spammer and get banned. Doesn't matter what the messages say.