The whole idea of post-scarcity economics is that there is no longer any need for money. Think Banks' Culture or Star Trek's Federation. I'm not saying it could or would work, and don't really want to get into that discussion; too much potential for going sour and there's masses of it in the literature already! But the utopian principal of post-scarcity is that all old models of economics collapse as completely irrelevant.
The key thing to realize is that incorporating AI into this system would not have made any improvement. While humans have very redundant sensory systems that are self-adjusting and self-calibrating, automated systems rely on a very limited set of fixed sensors that need to be carefully configured for desired operation. Giving an AI an increased number of sensors only increases the training difficulty. Additionally, humans are quite adept at at identifying inconsistencies, ignoring bad information, and being able to avoid doing the wrong thing. Machine learning is focused on training the AI to do the right thing under specific circumstances. I don't know of and approach to allow it to identify misinformation and to avoid making matters worse; the assumption is that all data is correct.
Star Trek displays a communistic, utopian world because of two things, a clean, abundant power source, and the replicator. The latter getting rid of a division of labor, thus the need for money and the collapse social class systems. DS9 poked holes in Roddenberry's vision in the episode where Quark tells Nog that humans are wonderful, so long as all their creature comforts are working, take those away and they'll become as "blood thirsty as any Klingon".
There are a couple of documentaries that do go into this in some detail. The system dips the nose in response to low air speed (which would normally be the correct action...so that the aircraft gains speed and avoids a stall). However, if the sensor malfunctions and reads low air speed erroneously, then it still dips the nose and potentially flies the aircraft into the ground (which happened twice, of course). There were secondary factors involved. Why only one sensor in the design (no redundancy)? And was the aircraft design inherently less stable, making it harder to fly and necessitating the complex automation between pilot and machine? But I think the pilots' main complaint was that they had not been sufficiently informed about this aspect of the system and were therefore not able to understand exactly what was going on. If they were aware of it, then they would have been able to successfully bypass it. So, yes, the determination to allow pilots with existing 737 credentials fly the new plane was a factor too. When the fault was recreated in a simulator, many 737 pilots had trouble understanding what was going on and keeping control. So the pilots in the two crashes were exonerated I believe. Definitely not Boeing's finest hour.
I don't consider the aforementioned systems 'AI', they're computerized safety features build into modern aircraft to help prevent accidents, based on (specific) programing.
If we did move into a technological utopian society, it would no doubt be gradual and we'd have time to adapt. If one day we had a replicator the concept of some being 'privileged' would be a thing of the past. If we could all have what we wanted money would be meaningless, there would be no need for it. However, for us to enjoy our 21st Century life styles we need a division labor, jobs that most probably wouldn't want to do. You can either pay them, or employ communism, which as history has shown devolves into tyranny. As you said, all social systems have flaws inherent to them. The 21st Century is bound to be a 'rough ride', however, I think we're slowly edging in the right direction, at least in certain parts of the world.
I noticed that under research, Bearly is listed twice. That certainly reduces my concerns with AI replacing researchers.
I noticed that as well, but in my following post there are two more.
Remember, though, that the current Machine Learning approach to AI is a two phase process. There is a learning phase then a cut-over to a usage phase. As I recall, ChatGPT-3 was trained on data through 2019. Research engines, as well as other applications in advancing fields, will not be privy to current findings. Maybe AI engines will be able to generate study of studies type papers, but I do not see them being able to access current research and certainly not advancing the cut edge of knowledge or hypothesis.
Correct me if I'm wrong, however you stated that the Airbus (to use your example) is cable of an 'autoland', which in your own words is the 'auto pilot landing the aircraft'. In my comment I stated 'in certain situations', meaning so long as everything goes 'according to plan' (data is inputted correctly, everything's functioning properly and the airport is equipped appropriately), the auto pilot can land the airplane. However, I'm fully aware that that process has to be monitored by the flight crew.
