When a flower gets more sunlight from one side than from the other, it starts to turn to that side. Is that different from a person who walks somewhere because they like to? Most people I ask say "Well, the human can say 'I'". But a computer can also say "I". Where is the threshold to call it "self-awareness"?
In other words, would you apply this to any system, or only biological ones. Is a sunflower turning to the sunlight as "programmed" as the thermostat?
Although, there is a difference; while an organism is the one saying "I" and not the cells or tissues. In a social group it's an individual (or group of them) that says "We".
An organism is a colony of cells and in that colony only some cells are responsible for thinking / communication.
On the flip side (or maybe not) there's epiphenomenalism and the narrative theory of the self.
Personally I think there's more evidence for the view that consciousness is a post hoc justification for unconscious actions than there is for it's "consciousness all the way down".
To me they seem the same. When you say "3 stones make a trio of stones" you don't mean that a "trio" existed somewhere outside of the stones and now moved into the stones, do you? You just gave a name to this specific formation of a stone, another stone and yet another stone.
I don't think they're necessarily inconsistent (hence "or maybe not"). But I do think that one person is unlikely to believe both of them.
Surely 'self-awareness' is intended to be about reasoning that the output shaft turned because [...], that the turn was made towards sunlight because [...]. I think this is what you're getting at with the human being able to say 'I', but I don't think you can gloss over that in that case, because it's the whole point isn't it?
You say "reasoning" is needed. Ok, but how is that defined? Say you "ask" the cog why it turned by rattling it. Then it will rattle the other cog and, therefore, point you towards it. And you might say "Oh, I understand. You turned because the other cog pushed you". You might say that is just a mechanical mechanism. But if you look at human behavior close enough, it's just a mechanical mechanism too.
In my experience, when you try to define and prove it in a way that is not fuzzy, you come up with a definition that also matches the cog.