The academic and author discusses what to expect from the singularity, the need for AI self-correcting mechanisms, and what hope there is for superintelligence safeguarding democracy.