In that view, the humans inside the corporation are like individuals neurons of a larger brain. The whole is more than the sum of its parts, and its incentives and actions are often not aligned with those of its parts, i.e., the humans.

The brain/neuron analogy is not a perfect description, because in order to function towards is potentially anti-human goals, the corporation must essentially bribe a small number of humans who are critical to its function. But this only goes so far: CEOs and other executives are well compensated, but they are easily replaced if they pose a serious threat to the goals of the corporation.

I think the argument is very compelling that today's AI Doomsday prophets are seriously missing the point, and we're actually already largely being ruled by hostile or at least indifferent non-human intelligences. It's the old sci-fi trope that if an intelligence is too alien, you might not even recognize it as one.