W Knight, et al, the original question is at the heart of many safety issues; it’s not just the FD, but human - system interaction and the wider issues of system design, training, checking, procedure advice / mandate, etc.
It’s difficult to ignore salient guidance cues overlaying the primary reference instrument - fundamental aspect of human performance - a similar cognitive feature as with illusion, how do you ignore an illusion (only after you realise that you have been mislead? ).
For those who wish to switch the FD off, another SOP, more paper work, memory item, and another opportunity to ‘fail’; how often do SOP writers ‘fail’.
It is impossible to expect every pilot to remember to take the correct action in every situation, particularly with the surprise of stall warning or stick push. Evidence from many accidents suggest that humans forget when surprised, cross crew monitoring and communication break down. ‘Pilot did not follow SOPs’, ‘did not recover from the stall using the correct procedure’, weak CRM, etc.
Many of modern day safety issues challenge the concept and limits of human performance; we cannot know and recall the ‘right’ thing to do in every situation.
The industry requires a different approach for these issues, but change is far from easy.
For those who favour automation, when do you remove the FD. Most aircraft use the FD for Windshear recovery which might infringe stick shake and then you inhibit valuable guidance.
Alternatively it should be possible to provide FD guidance for stall recovery, but would the effort of design, test, certification, not for all stall / upset scenarios, be cost effective if compared with efforts to avoid stall situations.
And as much as I favour spreading problem to a wider audience - management, regulator, manufacturer; action depends on their perception of the issue (or willingness to ‘see’), their view of safety priorities, or if there is a viable cost effective solution; so nothing happens.
On the other hand there may be great satisfaction in passing the problem on, being able to ‘blame’ someone else, “oh I told you so” after an event. cf 777 Asiana autothrust, 737 THY Rad Alt/autothrust, and as many examples as you wish to find or refute depending on your viewpoint.
“Wicked” problems can't be solved, but they can be tamed; how, recognise that there is a problem and be prepared to talk about it.