The pedagogical hard problem of generative AI: Socratic countermeasures

David J. Blacker states that generative AI has created a ‘pedagogical hard problem,’ forcing educators to verify genuine student experiences over final products. Referencing Henri Bergson, he argues that traditional text-based tasks no longer pass a ‘reverse Turing test’ (RTT), which involves the proving of human authorship. Blacker addresses this situation by offering an illustration of an RTT-passing performative, embodied learning activity, a student-designed AI in the form of a Socrates chatbot that can catalyze gradable philosophical performances resistant to AI outsourcing. Inspired by Socratic dialogue, this approach emphasizes experiential elements such as aporia, elenchus, parrhesia, and eunoia, which are best learned firsthand. Despite concerns about scalability and writing skills, Blacker believes that this approach can help sustain humanistic education amid the growth of AI.