Closed ML seems like a more apt description.

I don’t think intelligence is simply an inference engine or stochastic parrot, so in that sense, I think we’re far from the intelligence piece.

The attempt to unbias it is obviously a bias and that’s where the real issue with the closed-ness comes in.

I actually don’t mind a world in which we have competing (closed) models with their own biases, but that each make clear what they’re biasing, instead of pretending they’re not.

I’m such a context, people can choose what they want to use and for what purpose..