A learner model is a system’s internal profile of what a student likely knows, does not know, or may misunderstand. In educational AI, it is built from answers, explanations, timing, errors, and other behavior signals so the system can adapt exercises and feedback.
In cyber security, learner models matter because they contain sensitive behavioral data and can be manipulated. If an attacker injects misleading inputs, the system may infer the wrong skill level and give poor recommendations. If the model is exposed, it can reveal personal performance patterns about a child or employee. Defenses include data minimization, strict access control, retention limits, validation against human judgment, and monitoring for inconsistent or adversarial inputs. A good learner model should support decision-making without becoming a hidden surveillance or spoofing target.