This class pursues a twofold goal of making students both think about the use of AI in recidivism prediction/social benefits allocation and reflect on the possible use of AI for making evaluative, moral judgments. To do so, students are first presented with contextual information about recidivism prediction and social benefits allocation controversies before the class through mandatory reading. During the class, students are divided into four groups that take contrasting positions on these two issues (recidivism prediction and social benefits allocation) and are asked to suggest (counter-)arguments for the position that they have been assigned in a debate-style manner, following academic argumentative norms. Later, students use ChatGPT/GPT3 or an alternative LLM with a chat interface to prompt it about arguments and counter-arguments pertaining to their position. As a result, students compare and discuss the answers they have come up with and those offered by ChatGPT, both within the groups and with the whole class, exploring different dimensions of using AI for evaluative moral judgments