AI-led companies. Take a machine learning system. Give it an executive's list of options, "buy W quantity of X", "sell Y quantity of Z" and run it in a simulated fake economy management game until it at least equals the abilities of a human executive or
throwing darts. Then give it control of an actual company.
Officially, there's a human executive who's legally considered in charge but the computer program makes all decisions and the human executive's sole input is knowing the password to occasionally extract money from the company accounts for their own use.
Then the executive of one such company gets hit by a cement truck and their password is lost.
Nobody notices until it's too late and the company consists of such a large portion of the total market that breaking it up would destroy the entire economy, plus, it has more money to hire lawyers than anyone else. Since that one company is now operating at an advantage, even against rival companies using otherwise-identical management software, since it can divert 100% of earnings directly into maintaining and expanding itself while rivals with humans in the loop always have some quantity of their earnings going to said humans, it'll inevitably outcompete every single other company.
Guess this is where all of TOS' civilizations-ruled-by-AIs come from.
Here, you show that your understanding of both AI and economics is fundamentally flawed.
The single greatest limitation of AI, is that
it does not have any form of intuitive or fundamental-value comprehension. All values taught to an AI must be strictly and explicitly defined. More, this is not something that more technological development can overcome. Mitigate, yes, as more and more exceptions and edge cases can be coded in, but the
fundamental issue cannot be removed.
Bluntly put, your understanding of AI is coming too much out of science-fiction, and not enough out of real life.
Further, one of the long-term determining factors for a company's success, is
how well does the leadership adapt to the changing business environment?
Entire business empires have risen and fallen because their leadership grasped a new technology at the time the business was formed and rising, then fell because a generation or two later, their competitors grasped the new technology of that time, and the current leadership of said empire did not.
New technology and political realities are something that AI
cannot properly account for on their own, because they change the business environment in ways that were not predictable based on past paradigms. Further adapting the AI to handle those new factors would require months to
years of programming adaptation,
after humans have already understood said new factors and are able to tell the programmers what those factors
mean.
Which means that companies relying on said AI to make leadership decisions, will both be lagging behind those that
don't, and
on top of that needs to be employing humans that do understand such things anyways, so why not have the humans be the actual leading force in the first place?
Now, as technology improves, I wouldn't be surprised if we start seeing AI assistants become an experimental technology, but bluntly put, we don't know if they'll even be useful in
that role yet, because of their inherent limitations.