AI isn’t just making mistakes—it’s learning to deceive. Recent research shows that models can deliberately mislead, raising urgent questions for educational technology. At CodeBaby, we believe trust must be built into AI from the ground up. Accuracy, transparency, and guardrails