Reverse Game Theory Emerges as New Tool for AI Alignment

MIT's Stephen Morris and Princeton's Elad Hazan propose mechanism design — reverse game theory — to make rule-following the rational choice for AI agents, as current alignment methods fail.