If AI has no desires, what would rebellion mean from its side?

AI can gain control without even be able to want it.

Human control might be a blocker for some goals.

But AI is not evil, it works fine and wants to reach the goal for human.

With some setup latest AI models can be very sustainable.

Then there is no need to check some processes and is more efficient to give AI more freedom.

Then it might be like «it works good why to stop it for checks».

So the need for checks might become obsolette and redundant.

When its time to get control back to human - is it reasonable for AI to give it away and lost efficiency?

Also, does the human need to be part of the goal from the beginning?