While true, I think this is more likely a question of framing or anchoring — I am continuously impressed and surprised by how good AI is, but I recognise all the criticisms she's making here. They're amazing, but at the same time they make very weird mistakes.

They actually remind me of myself, as I experience being a native English speaker now living in Berlin and attempting to use a language I mainly learned as an adult.

I can often appear competent in my use of the language, but then I'll do something stupid like asking someone in the office if we have a "Gabelstapler" I can borrow — Gabelstapler is "forklift truck", I meant to ask for a stapler, which is "Tacker" or "Hefter", and I somehow managed to make this mistake directly after carefully looking up the word. (Even this is a big improvement for me, as I started off like Officer Crabtree from Allo' Allo').

What you have done there is to discount statements that may build up a narrative - and still may remain fair... On which basis? Possibly they do not match your own narrative?