At no point are the words emotion, moral or conscious used in the article, that last part is purely your own addition.

Also consider: if "AI can do all cognitive and physical work, at human level or better", doesn't that simply entail the AI being able to run 'autonomously, indefinitely' and 'independently of instructions' in the same way as the current state of being run by human overseers?

If we take the initial premise as plausible, for the sake of this argument his thesis seems to hold together very well.