Humans are very much inferior for tasks like these. There's a bunch of cases of pilots crashing the plane despite automation trying to save the plane. In fact that is far more common than automation crashing the plane.
I have an experience that relates to this.
Actually there are far more incidents that accidents, that are not in the media and even the passengers of those airplanes are not even aware that something happened. I had to analyse many of them and I will always remember my first one.
They brought me the FDR and CVR, the "black boxes" that are actually orange and not black. Nobody knew exactly what happened as it was my role to find out. The first to analyse is the Flight Data Recorder as you don't want your judgment to be influenced by pilot and co-pilot understanding of the situation as recorded in the Cockpit Voice Recorder.
Quick parameter analysis
To make it short, I found that while ground speed (Airplane speed compared to the ground) at one point decreased gradually, the air speed (Airplane speed compared to the surrounding air) suddenly spiked just before to reach a critical level then lower quite rapidly. At the same, altitude increased suddenly first then even faster few seconds after.
So what happened?
Airplane entered a air hole consequently air speed reached critical over-speed. To reduce speed, airplane has to go up (and not reduce gas), so the computer reacted immediately and the pilot had the same reaction just after. Hence because of this double reaction, correction became excessive and airplane went above max altitude it was designed for. Nobody was injured but because of this incident, the aircraft was not allowed to fly. Structure department took my report and had to define whether there could be internal structure damage and where.
The computer had a correct reaction to a problem: over-speed, so as the pilot. But it's the interaction between the 2 that failed. Actually the pilot should have known (it's part of their training) that the computer would react that way and should have monitored the reaction instead of reacting himself.
We are going toward pilot-less airplanes but we are not yet ready technically and psychologically. So we have automation with pilots.
That mandatory interaction between the 2, nowadays cause of many incidents, is very complicate to manage due to the radical differences between the 2 "thinking processes". One can correct the other deficiency but in some cases it creates or worsen the situation as my simple example showed.
It's a difficult transition. We shouldn't resist against this technological progress because it's not perfect. Once we are ready technologically and psychologically to pilot-less airplanes, pilots will be only a source of unpredictability and errors that we will need to get rid of. But as for now, we can't. So efforts should be made on making this interaction between human and computer as error-less as possible.
And tricking the 737max computer on the real gravity center position (due to the excessive engines weight) by changing its attack angle is, to me, completely against the philosophy. It can be a huge errors generator as the 2 accidents and multiple incidents seem to indicate!
