This is an odd point to try to argue in the context of machines. Of course the behaviour isn't fully intentional, it's a car, it has no mens rea. If I get in an

This is an odd point to try to argue in the context of machines. Of course the behaviour isn't fully intentional, it's a car, it has no mens rea. If I get in an accident because my car has a predictable manufacturing flaw where if I shift into third, then first, then fourth, then first again, then fifth, all the wheels fall off, does the fact that the exact same thing would happen again when repeated in similar circumstances mean my crash was an intentional decision by the car rather than an accident? You do raise an interesting point about how these behaviours are repeatable and predictable rather than random oopsies, though. I think it's more defensible to use that to emphasise that every accident a driverless car has, unless caused by a human, is a manufacturing flaw. It may not be inaccurate to say that the Waymo car had an accident where it dragged a person along underneath it, but it is much more accurate to say that Waymo cars had a manufacturing flaw where they would drag people underneath them, and an accident was caused by this flaw.