I don't understand what you mean when you refer to a singular "trusted authority".
Users have to trust that the group as a whole does not partake in any malicious behavior.
But this level of trust is already existent with the development of the reference client?
No.
Users can choose to use new versions of the software, if at all. But they cannot choose whether or not to receive alert messages.
Current situation is a good compromise; you're wrong, the alert system still exists, but it's determined by the software, and has a narrower range of events for which alerts are sent.
The behaviour you're seeking could be added programmatically (i.e. without requiring human input), you could write the code and submit it as a pull request for review.