Quite so.
Meanings do change over time and colloquial usage often differs markedly from technical definitions. That's why serious discussions avoid colloquial usage and stick to strictly defined terms to avoid misunderstanding.
I see where you are coming from. All the same, I happen to think that in this case the most inclusive definition is also the simplest and the most useful and legitimate.
Why so? Mostly because belief in God is supposed to be significant, to imply some sort of sense of wonder, of purpose or of gratitude.
I don't think it is very accurate or very respectful to take such deep emotional content as a default to be assume until evidence in contrary is found.
I agree with the above posts to the extent that i did find that definition to be the most helpful and the simplest in many situations. Particularly when trying to convey to someone else that there are varying degrees to the atheism of someone.
In this thread, i think what's most useful is to actually include all the known legitimate definitions, rather than just pick one. Reason being that the thread is particularly exploring how people understand atheism. Using that, we can then judge how helpful is the notion in question (the notion being categorizing babies as atheists).
Putting this in mind, i then think it's also reasonable to conclude that:
1) Technically, it's not false to say that babies are atheists.
2) However, it's not at all helpful to do so.
The reason i conclude the second point is because like i said in previous posts, it doesn't actually help us in categorizing babies at all. It doesn't tell me anything about them. It also doesn't negate other similar notions made by different religions, where the notion works logically within the framework of that religion.
I think what's more helpful is to set that group (babies and those unaware) aside, as a separate group of their own.