Education AI applications address two structural challenges: the resource constraint (one teacher to many students limits personalized attention) and the measurement gap (high-stakes assessments provide infrequent snapshots rather than continuous learning signals). AI systems that can provide individualized feedback at scale, and surface early signals of disengagement or struggle, address both. The field draws on decades of research in intelligent tutoring systems (ITS), beginning with Bloom’s 1984 “2 Sigma Problem” demonstrating that one-on-one tutoring produces two standard deviations of improvement over classroom instruction. AI now makes approximations of that tutoring ratio achievable at population scale. Deployment in schools operates under student data privacy obligations: FERPA (US), COPPA for under-13 data (US), and GDPR (EU).