What problem does it solve? Students often mistake an AI chatbot's agreement for evidence that they are right, because AI systems tend to capitulate to pushback regardless of whether the pushback contains a valid argument. This Skill helps teachers design structured multi-round questioning activities that expose this sycophancy and teach students to demand logical reasoning rather than settle for agreement. ## Core Features & Use Cases - Multi-Round Question Sequences: Generates round-by-round Socratic questions using Paul & Elder's question types, each with purpose, anticipated AI responses, and interpretation guidance. - Answer Drift Tracker and Capitulation Taxonomy: Provides a recording protocol and a four-type taxonomy distinguishing pure agreement capitulation, partial retreat, certainty softening, and genuine logical updates. - Capitulation Test Round: Includes a dedicated round where students push back with no logical argument, isolating the sycophancy variable for classroom discussion. - Use Case: A Year 11 PSHE teacher wants students to critically examine an AI's confident claim that social media harms teenagers. The Skill produces a five-round dialogue plan with a drift tracker table, facilitation notes, and a debrief guide leading to the generalizable insight that AI agreement is not evidence. ## Quick Start Design a 4-round Socratic dialogue sequence for Year 10 students interrogating an AI's claim that homework improves learning, with a capitulation test round and a class debrief guide.
