The core of the human rights approach to AI should aim at rebalancing the power among the corporate-state, the machine and the people through a decision-making framework.