In both these cases of politics immanent to autonomous art, it is because of these imagings of freedom and happiness that autonomous art becomes affirmative in the bad Marcusian sense of affirmative, and is viewed as such from the 1920s onwards. It is because autonomous art became affirmative of existing society at the same time as criticising it, because it was figuring the possibility of freedom and happiness in an unfree, unhappy society.
This complicated the critical criterion for the achievement of autonomy, adding an additional critical requirement for it, to counteract this affirmative function of autonomy; the requirement of an element of anti-art - an element of non-autonomous anti-art - within autonomous work, in order to render the autonomous work autonomous from its own affirmative autonomy.
It is only the anti-art element in twentieth-century art that marks the illusory character of the autonomy of the work. Since the artwork is autonomous by virtue of generating the illusion of autonomy, the problem is that people take the illusion of autonomy seriously. But it is not meant to be 'genuinely’ illusory, but self-consciously illusory.
What is important is that the artwork isn't 'actually' ontologically autonomous in some full sense, but rather 'appears' to be autonomous. The autonomous artwork is autonomous to the extent to which it can generate out of itself the appearance the illusion of its own autonomy.