Huy, P. Beyond the AI Feedback Paradox: How Metacognitive Re-Flection Transforms AI Corrections into L2 Learning. RSELTL 2026, 4, 823-846.