This paper presents a reflection on a master class on “Responsible Design of AI” aimed at raising critical thinking among students about the pros and cons of AI technology in everyday life usage on the example of chatbots. In contrast to typical approaches teaching existing policies and design guidelines, we aimed to challenge students by setting up a project on the “most unethical chatbot imaginable”. Our teaching concept therefore builds on students’ self-identified issues and concerns and develops guidelines for ethical chatbot design according to students’ interpretations of the capabilities and potential applications of these technologies. In our teaching we particularly focused on supporting mutual learning between teachers, students, and experts as foundational aspects. We conclude with reflections from the students regarding how this teaching approach can contribute to establish a critical and reflective mindset for future HCI researchers and developers.