Customer-Centric AI: the Major Paradigm Shift in AI Governance (Part 1) — LessWrong