So suppose Carol clicks "Contact Me Immediately Please" on a website and and enters her phone number... But--oops--there's a typo. Now Alice is going to get an "unsolicited" message even though literally everybody involved is operating in good faith.
Even if someone is maliciously pretending to be Alice, neither the website nor the phone-carrier has a better malice-detecting tool than simply sending it and seeing if the recipient replies "STOP".