Ali Kazad introduced an “Ethical AI Literacy” routine for B2 English learners using AI. Over six weeks, students performed scenario-based tasks, querying LLMs, then evaluating outputs for bias using an “Source, Bias, Impact” (SBI) Q-card, and revising language. While quick bias detection test scores remained flat, student reflections significantly deepened. They became much better at precisely naming biases, explaining necessary linguistic repairs, and planning future ethical AI use. The program aims to teach judgment for responsible AI integration, not ban tools, with an open-licensed toolkit available for teachers.