Incentive Compatibility for AI Alignment in Sociotechnical Systems: Positions and Prospects