AI doesn’t stall because models are weak; it stalls because people feel exposed. That is the central message from a new global report by Infosys and MIT Technology Review Insights: 83% of leaders say psychological safety measurably affects AI success, and 84% link it to business outcomes.