They are truly, as that one book title went, unsafe at any speed.
It's possible to get some kind of tactile response mimicking the edges of buttons well enough. I'd say a good implementation of this technology would be safe IF the position and availability of the controls would be static. I'm not sure what you'd get out of a touch screen with those constraints (less parts?), especially with the added cost and research, but that's the only way I can see a touch screen in a car that should ever be legal.
A Tesla driver in Germany already got fined for using the touch screen of his car while driving. It's strange how this obvious flaw of a car design has ever been allowed onto the roads.
Agreed -- this was also why the Apple Touchbar failed, IMO.
Typically keyboard use relies on a lot of muscle memory, and so the touchbar's "you get a different set of controls with different shapes and sizes that do different things depending on which app you're using" design ethos didn't really fit with how people actually use keyboards.
If buttons are too expensive and you want to reduce the count, if I press radio and then F1, that could be radio preset 1, and if I press fan and then F1, that turns the fan speed to 1 of 6 (which could be off, I don't really care), that would be fine, and then you don't need quite so many buttons. Otoh, if I'm paying $20k or more for a car, you can afford some buttons.
It's strange that anyone thinks it's even desirable. Touchscreens aren't even pleasant to use in this context.
When phones can run at 30 fps and have no noticeable lag, it's infuriating to switch to a car which stutters on some ugly, unnecessary graphics and takes 0.25 seconds to respond to a touch input. Not to mention the whole "you're supposed to be driving a car" aspect.