Our papers “Provably Optimal Reinforcement Learning under Safety Filtering” and “From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrails” were accepted to The International Association for Safe & Ethical AI (IASEAI’26).