the Schelling point for the caring threshold seems to be shifting toward progressively less intelligent (but still cute and harmless) species as time passes

"Cute" I'll give you.
"Harmless" I'm not sure about.

That is, it's not in the least bit clear to me that I can reliably predict, from species S being harmful and cute, that the Schelling point you describe won't/hasn't shifted so as to include S on the cared-about side.

For clarity: I make no moral claims here about any of this, and am uninterested in the associated moral claims, I'm just disagreeing with the bare empirical claim.