I think the fact that these symbols are "not words" is not in itself super earth-shattering. What really matters is whether they have to be learned like words. You could learn the words "New Driver" as a single "symbol", or you could learn the beginner symbol. The advantage of a verbal label is you can understand it without learning it, if you know the system its embdded in (i.e., English); its disadvantage is it's a big more complex and harder to distinguish from other symbols in that system if you don't know the system. The advantage of the symbol is it's compact and can be learned without having to take on the full complexity of a linguistic writing system; the disadvantage is it has to be learned separately by everyone, even those who already know a much more powerful system (namely, writing).
There are limits to both. Obviously it wouldn't be effective to have a bumper sticker that says "This driver has been driving for less than a year, so please use caution around them"; if you lean too much on the power of writing, you make it too hard to quickly identify the meaning. On the other hand, it's not easy to come up with a wide variety of sui generis symbols and still have them be visually distinguishable at a glance, so there's also a limit to how far that approach can get you.
I do wonder whether there is some link between differential reliance on symbols of this sort and the use of alphabetic vs. nonalphabetic writing systems in different languages/cultures. Like, crudely speaking, is it the case that, because you already have to learn so many separate symbols to read Japanese, learning a few more for stuff like "elderly driver" isn't much of an extra load? Or is it the reverse, that because you have to learn so many symbols to read Japanese, they want to make these separate symbols for a small set of important meanings so people can learn them even if they don't know how to read? I'm not sure whether there's a real link here or not but it's interesting to think about.