The reasons may not be as obviously correct as you think. It might, in theory, make sense to have 'smarter' people becoming doctors or engineers but this is not true for a couple of reasons. Firstly, the ATAR isn't even necessarily the perfect measure of someone's 'smartness'. Indeed, probably a better predictor of ATAR success is how rich your parents are rather than your IQ. You say a low ATAR tells its own 'story' and shows 'lack of commitment' but this is often very far from the truth. Here's a quick generalised example. Student A: Attends 5 different tutors for each subject, goes to a private school, has all the time in the world to study and has a plethora of resources at his/her disposal. Student B: Goes to a low SES school, has not even heard of external tutoring, works two jobs including night shifts sometimes just to help feed his/her family and barely gets time to study (or even a quiet place to do so). I'll let you imagine the difference in their outcomes. Another factor is that the entry requirements for the courses you mentioned are very high. Leaving behind the ATAR measure, are we really saying that a 95 scorer isn't 'good enough' for such a course, unlike a 99 scorer? A few ATAR points?
Here's a fun study for you to read --> https://www.researchgate.net/publication/329408770_Selection_and_lottery_in_medical_school_admissions_who_gains_and_who_loses
In short, it showed there is often very little or non-existent differences in the future performance of medical students picked through a stratified lottery system compared to an entry score based one. Meanwhile, there were massive gains in the diversity of the students in turn benefitting society as a whole.
ETA: Probably doesn't belong here. Feel free to move this conversation to appropriate thread mods!
Very late to this conversation, but just want to ++++1 this. Plenty of research that underlines this (incl the article shared already of course):
This article looks specifically at medical admissions and the influence of SES and gender. It found that female gender and low SES saw students underrepresented, but that students from these backgrounds did not under perform their counterparts. Major weakness was that it focused on one medical school and on the UMAT, which has since changed.
This article looks at the relationship between SES and ENTER scores (ATAR replaced them in 2011). It finds a relationship, but again that students who are disadvantaged are not actually less talented.
This one essentially found the same, but also found that disadvantaged students are no less likely to go to uni than their advantaged counterparts given the same ENTER scores.